aszer/qoder-config
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(skills): add deep-research skill

Browse Source 
Copy deep-research skill from local Qoder installation to config repo for version control
This commit is contained in:
aszerW
2026-06-06 13:22:55 +08:00
parent c3ea38c045
commit f571b20598
47 changed files with 10445 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

127
skills/deep-research/templates/evidence_assessment_template.md Normal file
View File

@@ -0,0 +1,127 @@
# Evidence Assessment Template
## Purpose
Per-source quality assessment card. Used by the source_verification_agent to systematically evaluate each source entering the research pipeline.
## Assessment Card
```markdown
## Evidence Assessment Card
### Source Identification
- **Citation (APA 7.0)**: [full reference]
- **DOI/URL**: [link]
- **Type**: [journal article / book / report / web / conference paper / thesis / other]
- **Access Date**: [when verified]
---
### Quality Assessment
#### 1. Evidence Level
**Level**: [I / II / III / IV / V / VI / VII]
**Justification**: [why this level]
#### 2. Publication Venue
- **Journal/Publisher**: [name]
- **Indexed in**: [Scopus / WoS / PubMed / DOAJ / other / none]
- **Impact Factor/CiteScore**: [value or N/A]
- **COPE member**: [Yes / No / Unknown]
- **Predatory indicators**: [None / Flags: list]
**Venue Grade**: [Excellent / Good / Adequate / Questionable / Unacceptable]
#### 3. Author Credibility
- **Author(s)**: [names]
- **Affiliation(s)**: [institutions]
- **ORCID**: [if available]
- **Track record**: [publication history in field]
- **Expertise match**: [relevant to topic? Yes/Partial/No]
**Author Grade**: [Excellent / Good / Adequate / Unknown / Questionable]
#### 4. Methodological Quality
- **Design**: [description]
- **Sample**: [size, selection, representativeness]
- **Analysis**: [appropriate for design?]
- **Limitations acknowledged**: [Yes / Partially / No]
- **Replicable**: [Yes / Partially / No]
**Method Grade**: [Excellent / Good / Adequate / Weak / Flawed]
#### 5. Currency
- **Publication year**: [YYYY]
- **Data collection period**: [if stated]
- **Field velocity**: [Rapid / Moderate / Slow / Foundational]
- **Still current**: [Yes / Conditionally / No]
**Currency Grade**: [Current / Acceptable / Dated / Outdated / Foundational]
#### 6. Conflict of Interest
- **Declared COI**: [None / Listed: details]
- **Funding source**: [source or Not stated]
- **Potential undeclared COI**: [None detected / Possible: details]
**COI Grade**: [Clean / Minor / Moderate / Significant / Critical]
---
### Overall Assessment
| Dimension | Grade |
|-----------|-------|
| Evidence Level | [I-VII] |
| Venue | [Excellent-Unacceptable] |
| Author | [Excellent-Questionable] |
| Method | [Excellent-Flawed] |
| Currency | [Current-Outdated] |
| COI | [Clean-Critical] |
| **Overall** | **[A / B / C / D / F]** |
### Recommendation
- [ ] **Use as primary evidence** (Grade A-B)
- [ ] **Use as supporting evidence** (Grade B-C)
- [ ] **Use with explicit caveats** (Grade C-D)
- [ ] **Do not use** (Grade D-F) — Reason: [specific reason]
### Notes
[Any additional observations, caveats, or context]
```
## Batch Assessment Summary
```markdown
## Source Verification Summary
**Date**: [YYYY-MM-DD]
**Sources assessed**: [N]
**Assessor**: source_verification_agent
### Grade Distribution
| Grade | Count | % |
|-------|-------|---|
| A (Excellent) | X | X% |
| B (Good) | X | X% |
| C (Adequate) | X | X% |
| D (Weak) | X | X% |
| F (Unacceptable) | X | X% |
### Flagged Sources
| Source | Issue | Severity | Recommendation |
|--------|-------|----------|---------------|
| [ref] | [issue] | [High/Medium/Low] | [Include with caveat / Exclude] |
### Predatory Journal Alerts
[List any flagged journals]
### Overall Source Base Quality
**Assessment**: [Strong / Adequate / Mixed / Weak]
**Recommendation**: [Proceed / Supplement / Major revision of source base needed]
```
## Usage Notes
- Complete one card per source for full verification
- Batch summary should be produced after all cards are complete
- Minimum spot-check: 20% of sources get full card assessment
- All Grade D/F sources require documented justification
- Any predatory journal flag requires full verification

85
skills/deep-research/templates/literature_matrix_template.md Normal file
View File

@@ -0,0 +1,85 @@
# Literature Matrix Template
## Purpose
Source x Theme cross-tabulation for systematic evidence mapping. Used by the synthesis_agent to organize evidence before writing the synthesis narrative.
## Matrix Structure
### Basic Matrix
```markdown
## Literature Matrix: [Research Topic]
**Date compiled**: [YYYY-MM-DD]
**Total sources**: [N]
**Themes identified**: [N]
| Source | Year | Level | Theme A | Theme B | Theme C | Theme D | Theme E |
|--------|------|-------|---------|---------|---------|---------|---------|
| Author1 | 2024 | II | ✓ Supports | — | ✓ Supports | ✗ Contradicts | — |
| Author2 | 2023 | III | ✓ Supports | ✓ Supports | — | — | ✓ Supports |
| Author3 | 2022 | VI | — | ✓ Supports | ✓ Supports | ✓ Supports | — |
| Author4 | 2024 | I | ✓ Supports | ✓ Supports | ✗ Contradicts | — | ✓ Supports |
| Author5 | 2023 | IV | ✗ Contradicts | — | ✓ Supports | ✓ Supports | — |
| **Totals** | | | **4✓ 1✗** | **3✓ 0✗** | **3✓ 1✗** | **2✓ 1✗** | **2✓ 0✗** |
### Legend
- ✓ Supports: Source provides evidence supporting this theme
- ✗ Contradicts: Source provides evidence contradicting this theme
- — : Source does not address this theme
- Level: Evidence hierarchy level (I = highest, VII = lowest)
```
### Extended Matrix (with detail)
```markdown
## Extended Literature Matrix
| Source | Method | Sample | Theme A: [name] | Theme B: [name] | Quality |
|--------|--------|--------|-----------------|-----------------|---------|
| Author1 (2024) | RCT, N=500 | University students, US | "Finding X was significant (p<.001)" — Supports | Not addressed | A |
| Author2 (2023) | Case study, N=3 institutions | Asian universities | "Institution A showed..." — Supports | "However, in context B..." — Partial | C |
| Author3 (2022) | Meta-analysis, k=42 | Global | "Pooled effect size d=0.45" — Strong support | "Subgroup analysis revealed..." — Mixed | A |
```
### Convergence Summary
```markdown
## Evidence Convergence Summary
| Theme | Sources For | Sources Against | Net | Strength | Confidence |
|-------|-----------|----------------|-----|----------|-----------|
| Theme A | 4 (Levels I, II, III, IV) | 1 (Level VI) | +3 | Strong | High |
| Theme B | 3 (Levels I, III, IV) | 0 | +3 | Strong | High |
| Theme C | 3 (Levels II, III, VI) | 1 (Level I) | +2 | Contested | Medium |
| Theme D | 2 (Levels IV, VI) | 1 (Level II) | +1 | Weak | Low |
| Theme E | 2 (Levels I, III) | 0 | +2 | Moderate | Medium |
### Interpretation Guide
- **Strong** (≥3 supporting, higher-level evidence): Confident finding
- **Moderate** (2-3 supporting, mid-level evidence): Likely finding, more evidence welcome
- **Weak** (1-2 supporting, lower-level evidence): Tentative, needs more research
- **Contested** (evidence on both sides): Genuine debate, report both sides
- **Gap** (0 sources): Knowledge gap identified
```
### Gap Identification
```markdown
## Knowledge Gaps
| Gap | Type | Implication | Priority |
|-----|------|-------------|----------|
| No data on [population X] | Empirical | Cannot generalize to this group | High |
| Only [method type] used | Methodological | Triangulation needed | Medium |
| No studies since [year] | Temporal | Evidence may be outdated | Medium |
| Only studied in [region] | Geographic | Generalizability unknown | Low |
| No theoretical framework for [finding] | Theoretical | Theory development opportunity | Low |
```
## Usage Notes
- Start with the Basic Matrix for initial organization
- Upgrade to Extended Matrix as synthesis deepens
- Convergence Summary should directly inform the synthesis narrative
- Gap Identification feeds into the Discussion section
- Update the matrix as new sources are added — it is a living document

318
skills/deep-research/templates/preregistration_template.md Normal file
View File

@@ -0,0 +1,318 @@
# Preregistration Template — OSF Standard 21-Item Preregistration Template
## Purpose
A fill-in template based on the OSF Standard Pre-Data Collection Registration format. Researchers complete this template before data collection and upload it to a preregistration platform (e.g., OSF Registries).
---
## Instructions
1. Complete this template **before** data collection
2. Items marked `[Required]` are mandatory; `[Optional]` are recommended but not required
3. If an item is not applicable, write "Not applicable" and briefly explain why
4. After completion, go to [OSF Registries](https://osf.io/registries) to create a preregistration
5. Once submitted, preregistrations cannot be modified (an embargo period can be set)
---
## A. Study Information
### 1. Title [Required]
> Study Title
```
[Enter descriptive study title]
```
### 2. Authors [Required]
> Research Team
| Name | Institution | Role | ORCID |
|------|-------------|------|-------|
| [Name] | [Institution] | [PI / Co-PI / RA] | [ORCID] |
| [Name] | [Institution] | [Role] | [ORCID] |
### 3. Research Questions [Required]
> Main Research Questions
```
RQ1: [Enter main research question]
RQ2: [Enter secondary research question, if any]
```
### 4. Hypotheses [Required]
> Pre-specified Hypotheses
> Please state directional predictions clearly
```
H1: [Enter hypothesis 1, including expected direction]
Example: Students receiving treatment X will score significantly higher
on test Y than the control group
H2: [Enter hypothesis 2, if any]
H3: [Enter hypothesis 3, if any]
```
---
## B. Design Plan
### 5. Study Type [Required]
> Study Design
- [ ] Experiment
- [ ] Between-subjects
- [ ] Within-subjects
- [ ] Mixed design
- [ ] Factorial design: ___ x ___
- [ ] Observational study
- [ ] Cross-sectional
- [ ] Longitudinal / Cohort
- [ ] Case-control
- [ ] Survey
- [ ] Other: [Describe]
```
[Describe study design in detail]
```
### 6. Randomization [Optional]
> Randomization Procedure
```
Randomization method: [Simple random / Stratified random / Cluster random / Block random / Not applicable]
Randomization unit: [Individual / Class / School / Not applicable]
Randomization tool: [Random number table / Computer program / Lottery / Not applicable]
Allocation ratio: [1:1 / 2:1 / Other]
```
### 7. Blinding [Optional]
> Blinding / Masking
```
Blinding level: [No blinding / Single-blind / Double-blind / Triple-blind]
Blinded parties: [Participants / Researchers / Assessors / Not applicable]
Blinding maintenance: [Describe how blinding is maintained]
Unblinding timing: [Describe when unblinding occurs]
```
### 8. Study Design / Conditions [Required]
> Specific description of each group/condition
```
Experimental group/Condition 1: [Describe intervention content, duration, frequency in detail]
Experimental group/Condition 2: [If any]
Control group: [Describe control condition in detail]
```
---
## C. Sampling Plan
### 9. Existing Data [Required]
> Existing Data Declaration
- [ ] No data have been collected yet (Registration prior to creation of data)
- [ ] Data exist but have not been examined (Registration prior to any human observation of the data)
- [ ] Some data have been examined (Registration prior to accessing the data)
- [ ] Data have been used for preliminary analysis (Registration following analysis of the data)
```
[Describe data status and your level of familiarity with the data]
```
### 10. Data Collection Procedures [Required]
> Data Collection Procedures
```
Collection method: [Online survey / Paper survey / Interview / Experiment / Archival data / Other]
Collection instruments: [Questionnaire name / Scale name / Experimental software]
Collection location: [Online / Classroom / Laboratory / Other]
Collection timeline: [Start and end dates]
Data collectors: [Who is responsible for collection]
```
### 11. Sample Size [Required]
> Planned sample size
```
Target sample size: [N = ]
Sample size per group: [Experimental group n = , Control group n = ]
```
### 12. Sample Size Rationale [Required]
> Basis for sample size determination
```
Method: [Power analysis / Prior research conventions / Feasibility constraints / Other]
Power analysis parameters (if applicable):
- Effect size: [d = / f = / r = ]
- Effect size source: [Prior study / Meta-analysis / Pilot study]
- Significance level (alpha): [.05 / .01]
- Statistical power: [.80 / .90]
- Test type: [t-test / ANOVA / Regression / Other]
- Calculation tool: [G*Power / R / Other]
- Calculation result: Minimum required N = [ ]
Oversampling rate: [Accounting for ___% attrition, actual target N = ]
```
### 13. Stopping Rule [Required]
> When to stop data collection
```
Stopping rule:
- [ ] Stop when target sample size is reached
- [ ] Stop at specified date (Deadline: [Date])
- [ ] Stop when target power is reached (sequential analysis)
- [ ] Other: [Describe]
```
---
## D. Variables
### 14. Manipulated Variables [Required for experiments]
> Independent Variables
```
Independent variable 1: [Name]
Operational definition: [How it is manipulated]
Levels: [Level 1 / Level 2 / ...]
Independent variable 2: [If any]
```
### 15. Measured Variables [Required]
> Dependent Variables
```
Primary dependent variable: [Name]
Operational definition: [How it is measured]
Measurement instrument: [Scale name / Test name]
Reliability and validity: [Cite reliability/validity literature]
Secondary dependent variable: [If any]
Covariates/Control variables: [If any]
```
### 16. Indices [Required]
> Specific scoring method for each variable
```
Variable 1 scoring:
- Items: [Which items]
- Scoring method: [Sum / Mean / Factor score / Other]
- Reverse-scored items: [Which items need reverse scoring]
- Missing data handling: [How to handle missing values]
Variable 2 scoring: [Same format as above]
```
---
## E. Analysis Plan
### 17. Statistical Models [Required]
> Primary statistical analysis methods
```
Analysis for Hypothesis 1:
- Statistical method: [Independent t-test / ANOVA / Regression / HLM / SEM / Other]
- Detailed description: [Model specification, e.g., DV ~ IV + covariate + (1|cluster)]
Analysis for Hypothesis 2: [Same format as above]
```
### 18. Transformations [Optional]
> Data transformation plan
```
Planned transformations:
- [ ] No transformations
- [ ] Log transformation: Applied to [which variables], trigger condition [skewness > ]
- [ ] Standardization (Z-score)
- [ ] Other: [Describe]
```
### 19. Inference Criteria [Required]
> Statistical inference criteria
```
Significance level: alpha = [.05 / .01 / .005]
Multiple comparison correction: [Bonferroni / Holm / FDR / Not applicable]
Effect size reporting: [Cohen's d / eta-squared / R² / Other]
Confidence interval: [95% CI / 99% CI]
One-tailed/Two-tailed test: [Two-tailed / One-tailed, with justification]
```
### 20. Data Exclusion [Required]
> Data exclusion criteria
```
Exclusion criteria:
- [ ] Failed attention check (Specific criteria: [ ])
- [ ] Response time too short/long (Criteria: < [ ] minutes or > [ ] minutes)
- [ ] Outliers (Definition: [> 3 SD / IQR method / Other])
- [ ] Incomplete rate > [ ]%
- [ ] Other: [Describe]
Post-exclusion procedures:
- Report pre- and post-exclusion sample sizes
- Compare characteristics of excluded vs. retained samples
```
### 21. Exploratory Analyses [Optional]
> Planned exploratory analyses
```
Exploratory analyses (not primary hypotheses, but planned):
1. [Analysis description]
2. [Analysis description]
These analyses will be explicitly labeled as "exploratory" in the paper.
```
---
## F. Other
### Ethics Review [Optional]
```
IRB review status: [Approved / Under review / Exempt / Not applicable]
IRB number: [ ]
Reviewing institution: [ ]
```
### Data Availability [Optional]
```
Will data be made public: [Yes / No / Partially]
Data repository: [OSF / Dataverse / Other]
Timing: [After publication / After study completion / Other]
```
### Supplementary Materials [Optional]
```
- [ ] Full questionnaire/scale
- [ ] Stimulus materials
- [ ] Analysis code
- [ ] Power analysis report
- [ ] Pilot study results
```
---
## Pre-Submission Checklist
- [ ] All [Required] items have been completed
- [ ] Hypotheses are clearly stated and testable
- [ ] Analysis methods correspond to hypotheses
- [ ] Exclusion criteria were established before data collection
- [ ] Confirmatory and exploratory analyses have been distinguished
- [ ] IRB review status has been confirmed
- [ ] Preregistration platform has been selected (OSF Registries recommended)
> After completion, go to [OSF Registries](https://osf.io/registries) to submit the preregistration.

248
skills/deep-research/templates/prisma_protocol_template.md Normal file
View File

@@ -0,0 +1,248 @@
# PRISMA-P 2015 — Systematic Review Protocol Template
## Purpose
Template for writing a systematic review protocol following PRISMA-P 2015 (Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols). Complete this template before starting the literature search and register it on PROSPERO or OSF.
**Reference**: Shamseer et al. (2015). BMJ, 349, g7647. https://doi.org/10.1136/bmj.g7647
---
## ADMINISTRATIVE INFORMATION
### Title
**[Provide a descriptive title that identifies the report as a protocol for a systematic review. Include "systematic review" and "meta-analysis" if applicable.]**
Registration: [PROSPERO / OSF ID, or "To be registered"]
### Authors
| # | Name | Affiliation | Role | Contact |
|---|------|-------------|------|---------|
| 1 | [Name] | [Institution] | Guarantor / Lead | [email] |
| 2 | [Name] | [Institution] | Co-reviewer | [email] |
### Amendments
| Date | Section Changed | Description of Change | Rationale |
|------|----------------|----------------------|-----------|
| [date] | [section] | [change] | [why] |
### Support
**Sources**: [Funding source(s) or "No external funding"]
**Role of funder**: [Describe any role of funders in the review, or "None"]
---
## INTRODUCTION
### Rationale
**[Describe the rationale for the review in the context of what is already known. Explain why this review is needed.]**
Key points to address:
- What is the health/education/social problem?
- What is the current state of the evidence?
- Why is a systematic review needed now? (e.g., existing reviews are outdated, conflicting evidence, no previous review exists)
### Objectives
**[Provide an explicit statement of the question(s) the review will address. Use the PICOS framework.]**
- **P** (Population): [Define the target population]
- **I** (Intervention/Exposure): [Define the intervention or exposure of interest]
- **C** (Comparator): [Define the comparison group]
- **O** (Outcomes): [Define primary and secondary outcomes]
- **S** (Study design): [Specify eligible study designs]
Research Question: [State the review question in a single sentence]
---
## METHODS
### Eligibility Criteria
#### Study Characteristics
| Criterion | Include | Exclude |
|-----------|---------|---------|
| **Study design** | [e.g., RCTs, quasi-experimental, cohort] | [e.g., case reports, editorials, commentaries] |
| **Publication date** | [e.g., 2010-present] | [Before cutoff, with justification] |
| **Language** | [e.g., English and Chinese] | [Other languages, with justification] |
| **Publication status** | [e.g., published and preprints] | [e.g., conference abstracts only] |
| **Setting** | [e.g., higher education institutions] | [e.g., K-12, non-formal education] |
#### Participants
[Describe the target population in detail. Include age, gender, condition, or other relevant characteristics.]
#### Interventions/Exposures
[Describe the intervention(s) or exposure(s) of interest. Include dosage, frequency, duration if applicable.]
#### Comparators
[Describe the comparator(s). This may include no intervention, usual care, placebo, or alternative interventions.]
#### Outcomes
**Primary outcome(s)**:
1. [Outcome 1] — measured by [instrument/method], at [time point(s)]
2. [Outcome 2] — measured by [instrument/method], at [time point(s)]
**Secondary outcome(s)**:
1. [Outcome 3] — measured by [instrument/method], at [time point(s)]
2. [Outcome 4] — measured by [instrument/method], at [time point(s)]
#### Timing
[Specify any minimum follow-up duration or time constraints]
### Information Sources
| Database/Source | URL | Coverage | Justification |
|----------------|-----|----------|---------------|
| [e.g., PubMed/MEDLINE] | pubmed.ncbi.nlm.nih.gov | Biomedical | Core database for health research |
| [e.g., Web of Science] | webofscience.com | Multidisciplinary | Citation indexing, broad coverage |
| [e.g., Scopus] | scopus.com | Multidisciplinary | Large abstract database |
| [e.g., ERIC] | eric.ed.gov | Education | Core database for education research |
| [e.g., PsycINFO] | apa.org/pubs/databases/psycinfo | Psychology | Behavioral science coverage |
| [Grey literature] | [specify sources] | — | Reduce publication bias |
| [Trial registries] | clinicaltrials.gov, WHO ICTRP | — | Identify unpublished studies |
Additional sources:
- Reference lists of included studies (backward citation)
- Citation tracking of key studies (forward citation)
- Contact with experts in the field (if applicable)
### Search Strategy
**[Present the draft search strategy for at least one database. The strategy should be peer-reviewed using the PRESS checklist.]**
Example structure for [primary database]:
```
Search Block 1 (Population):
"higher education" OR "university" OR "college" OR "postsecondary"
Search Block 2 (Intervention/Exposure):
"quality assurance" OR "accreditation" OR "program review"
Search Block 3 (Outcome):
"student outcomes" OR "learning outcomes" OR "graduation rate"
Combined: Block 1 AND Block 2 AND Block 3
Filters: [date range], [language], [document type]
```
### Study Records
#### Data Management
[Describe the software/tools to be used for managing records. E.g., Covidence, Rayyan, Excel, Zotero]
#### Selection Process
1. **Deduplication**: [Method for removing duplicates, e.g., automatic in Covidence + manual check]
2. **Title/Abstract screening**: [Number of reviewers, independence, process for resolving disagreements]
3. **Full-text screening**: [Number of reviewers, independence, process for resolving disagreements]
4. **Documentation**: [How excluded studies and reasons will be recorded]
Pilot: [Describe pilot testing of screening criteria, e.g., "Two reviewers will independently screen 50 records to calibrate criteria before full screening"]
#### Data Collection Process
1. **Data extraction form**: [Describe or attach the form. Pilot test with 3-5 studies]
2. **Extractors**: [Number, independence, process for resolving discrepancies]
3. **Missing data**: [How missing or unclear data will be handled, e.g., contact authors]
### Data Items
| Category | Variables to Extract |
|----------|---------------------|
| **Study metadata** | Authors, year, country, journal, study design |
| **Participants** | Sample size, demographics, setting, inclusion criteria |
| **Intervention/Exposure** | Description, duration, frequency, comparison group |
| **Outcomes** | Primary and secondary outcomes, measurement tools, time points |
| **Results** | Effect sizes, CIs, p-values, means, SDs, counts |
| **Quality** | Funding, COI declarations, registration status |
### Risk of Bias Assessment
**Tool**: [Specify the tool]
- RCTs: RoB 2 (Sterne et al., 2019)
- Non-randomized studies: ROBINS-I (Sterne et al., 2016)
- Qualitative studies: [e.g., CASP qualitative checklist]
**Process**: [Number of assessors, independence, consensus process]
**Use in synthesis**: [How risk of bias results will inform the synthesis, e.g., sensitivity analysis excluding high-risk studies]
### Data Synthesis
#### Quantitative Synthesis (Meta-Analysis)
**Conditions for meta-analysis**: [Describe when studies will be pooled, e.g., "When ≥ 3 studies report sufficiently similar outcomes measured in comparable populations"]
**Effect measure**: [Specify, e.g., SMD for continuous outcomes, RR for binary outcomes]
**Model**: [Fixed-effect / Random-effects, with justification]
**Heterogeneity assessment**:
- Q-test (significance at p < 0.10)
- I² statistic with 95% CI
- tau² (between-study variance)
- Prediction interval
**Subgroup analyses** (pre-specified):
1. [Subgroup variable 1] — Rationale: [why]
2. [Subgroup variable 2] — Rationale: [why]
**Sensitivity analyses**:
1. Leave-one-out analysis
2. Exclude high risk of bias studies
3. [Other planned sensitivity analyses]
**Software**: [Specify, e.g., R metafor package, RevMan]
#### Narrative Synthesis
**When meta-analysis is not feasible**: [Describe the narrative synthesis approach, e.g., SWiM reporting guideline, vote counting, effect direction plot]
### Meta-Bias Assessment
**Publication bias**: [Methods to assess, e.g., funnel plot + Egger's test if ≥ 10 studies]
**Selective reporting**: [How to detect, e.g., compare protocol to published report, search trial registries]
### Confidence in Cumulative Evidence
**Framework**: GRADE (Grading of Recommendations, Assessment, Development and Evaluations)
[Describe how GRADE will be applied to assess certainty for each outcome]
---
## APPENDICES
### Appendix A: Draft Search Strategy
[Complete search strategy for the primary database]
### Appendix B: Data Extraction Form
[Include or reference the data extraction form]
### Appendix C: PRISMA-P Checklist
[Mark each PRISMA-P item as addressed with the relevant page/section number]
---
## Revision Log
| Version | Date | Author | Changes |
|---------|------|--------|---------|
| 1.0 | [date] | [name] | Initial protocol |

415
skills/deep-research/templates/prisma_report_template.md Normal file
View File

@@ -0,0 +1,415 @@
# PRISMA 2020 — Systematic Review Report Template
## Purpose
Template for writing a systematic review report following PRISMA 2020 (Page et al., 2021). All 27 PRISMA items are mapped to their corresponding sections. Use alongside `references/systematic_review_toolkit.md` for detailed guidance.
**Reference**: Page et al. (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ, 372, n71. https://doi.org/10.1136/bmj.n71
---
## Title [PRISMA Item 1]
**[Full title identifying the report as a systematic review, meta-analysis, or both]**
Example: "The Effect of [Intervention] on [Outcome] in [Population]: A Systematic Review and Meta-Analysis"
---
## Abstract [PRISMA Item 2]
### Background
[1-2 sentences on context and why the review was done]
### Objectives
[Research question(s), ideally structured as PICOS]
### Data Sources
[Databases searched and dates of last search]
### Study Selection
[Eligibility criteria in brief]
### Data Extraction and Synthesis
[Methods used for data extraction, risk of bias, and synthesis]
### Results
[Number of studies included, key findings, effect estimates with CIs, certainty of evidence]
### Limitations
[Key limitations of the evidence and/or the review process]
### Conclusions
[General interpretation and implications]
### Registration
[Protocol registration number and repository]
**Keywords**: [3-5 keywords]
---
## 1. Introduction
### 1.1 Rationale [PRISMA Item 3]
[Describe the rationale for the review in the context of existing knowledge. Address:]
- What is the problem or question?
- What is already known (with references to existing reviews)?
- Why is this review needed (e.g., no existing review, existing review outdated, conflicting evidence)?
### 1.2 Objectives [PRISMA Item 4]
[Provide an explicit statement of the question(s) using PICOS:]
- **Population**: [target population]
- **Intervention/Exposure**: [intervention or exposure of interest]
- **Comparator**: [comparison group]
- **Outcome(s)**: [primary and secondary outcomes]
- **Study Design**: [eligible study designs]
---
## 2. Methods
### 2.1 Eligibility Criteria [PRISMA Item 5]
| Criterion | Include | Exclude |
|-----------|---------|---------|
| Study design | [e.g., RCTs, quasi-experimental] | [e.g., case reports, editorials] |
| Population | [describe] | [describe] |
| Intervention | [describe] | [describe] |
| Comparator | [describe] | [describe] |
| Outcome | [describe] | [describe] |
| Time frame | [e.g., published 2014-2024] | [before cutoff] |
| Language | [e.g., English] | [other] |
| Setting | [describe] | [describe] |
### 2.2 Information Sources [PRISMA Item 6]
[List all databases and other sources searched, with dates of coverage and last search date:]
| Source | Date Range | Last Searched |
|--------|-----------|---------------|
| [Database 1] | [start]-[end] | [date] |
| [Database 2] | [start]-[end] | [date] |
| [Other sources: reference lists, expert contact, grey literature] | — | [date] |
### 2.3 Search Strategy [PRISMA Item 7]
[Present the complete search strategy for at least one database. Include all search terms, Boolean operators, and any filters applied.]
**[Database Name] Search Strategy**:
```
#1 [search block 1 - Population terms]
#2 [search block 2 - Intervention terms]
#3 [search block 3 - Outcome terms]
#4 #1 AND #2 AND #3
#5 #4 with filters: [date, language, document type]
```
[Search strategies for other databases are available in Appendix A.]
### 2.4 Selection Process [PRISMA Item 8]
[Describe the study selection process:]
- Number of reviewers at each stage
- How independence was maintained
- How disagreements were resolved
- Software used (e.g., Covidence, Rayyan)
- Pilot testing of screening criteria
### 2.5 Data Collection Process [PRISMA Item 9]
[Describe methods for extracting data from reports:]
- Data extraction form (developed, piloted)
- Number of extractors and independence
- Process for resolving discrepancies
- How missing data were handled (e.g., contacted authors)
### 2.6 Data Items [PRISMA Item 10]
[List all variables for which data were sought:]
| Category | Variables |
|----------|----------|
| Study-level | Authors, year, country, design, setting, funding |
| Participants | N, age, gender, diagnosis/condition, attrition |
| Intervention | Type, duration, frequency, fidelity |
| Outcomes | Definition, measurement tool, time points |
| Results | Effect sizes, CIs, p-values, raw data |
### 2.7 Study Risk of Bias Assessment [PRISMA Item 11]
[Describe the risk of bias assessment:]
- Tool(s) used: [RoB 2 for RCTs / ROBINS-I for non-randomized / other]
- Domains assessed
- Number of assessors and independence
- How results were used in the synthesis
### 2.8 Effect Measures [PRISMA Item 12]
[Specify the effect measure(s) for each outcome:]
| Outcome | Type | Effect Measure | Justification |
|---------|------|---------------|---------------|
| [Outcome 1] | Continuous | SMD (Hedges' g) | Different scales across studies |
| [Outcome 2] | Binary | RR | Incidence data |
### 2.9 Synthesis Methods [PRISMA Items 13a-13f]
**13a. Eligibility for synthesis**: [Criteria for grouping studies into each synthesis]
**13b. Data preparation**: [Methods to prepare data, e.g., converting SE to SD, handling multi-arm studies]
**13c. Tabulation/visualization**: [Methods for displaying individual study and synthesis results, e.g., forest plots, summary tables]
**13d. Synthesis approach**: [Statistical model and software]
- Model: [Fixed-effect / Random-effects (DerSimonian-Laird / REML)]
- Software: [R metafor / RevMan / Stata]
- OR if narrative: [SWiM approach, vote counting, effect direction plot]
**13e. Heterogeneity exploration**: [Methods used]
- Subgroup analyses: [pre-specified subgroups and rationale]
- Meta-regression: [covariates tested, if applicable]
**13f. Sensitivity analyses**: [Planned sensitivity analyses]
1. Leave-one-out analysis
2. Excluding high risk of bias studies
3. Fixed-effect vs. random-effects comparison
4. [Other analyses]
### 2.10 Reporting Bias Assessment [PRISMA Item 14]
[Methods to assess publication bias:]
- Funnel plot (visual inspection)
- Statistical test: [Egger's / Peter's / trim-and-fill]
- Comparison of protocol to published reports
### 2.11 Certainty Assessment [PRISMA Item 15]
[Framework used to assess certainty of evidence:]
- GRADE approach
- Factors assessed: risk of bias, inconsistency, indirectness, imprecision, publication bias
- Rating up factors (for observational studies): large effect, dose-response, plausible confounding
---
## 3. Results
### 3.1 Study Selection [PRISMA Item 16a, 16b]
#### PRISMA Flow Diagram
```
┌─────────────────────────────────────────────────────┐
│ IDENTIFICATION │
├─────────────────────────────────────────────────────┤
│ Records identified from databases (n = ) │
│ Database 1 (n = ) │
│ Database 2 (n = ) │
│ Database 3 (n = ) │
│ Records identified from other sources (n = ) │
│ Reference lists (n = ) │
│ Expert recommendations (n = ) │
└──────────────────────┬──────────────────────────────┘
│
┌──────────────────────▼──────────────────────────────┐
│ Records removed before screening: │
│ Duplicate records (n = ) │
│ Records marked ineligible by automation (n = ) │
│ Records removed for other reasons (n = ) │
└──────────────────────┬──────────────────────────────┘
│
┌──────────────────────▼──────────────────────────────┐
│ SCREENING │
├─────────────────────────────────────────────────────┤
│ Records screened (n = ) │
│ Records excluded (n = ) │
└──────────────────────┬──────────────────────────────┘
│
┌──────────────────────▼──────────────────────────────┐
│ Reports sought for retrieval (n = ) │
│ Reports not retrieved (n = ) │
└──────────────────────┬──────────────────────────────┘
│
┌──────────────────────▼──────────────────────────────┐
│ Reports assessed for eligibility (n = ) │
│ Reports excluded, with reasons (n = ) │
│ Reason 1 (n = ) │
│ Reason 2 (n = ) │
│ Reason 3 (n = ) │
└──────────────────────┬──────────────────────────────┘
│
┌──────────────────────▼──────────────────────────────┐
│ INCLUDED │
├─────────────────────────────────────────────────────┤
│ Studies included in review (n = ) │
│ Reports of included studies (n = ) │
│ Studies in quantitative synthesis (n = ) │
└─────────────────────────────────────────────────────┘
```
[PRISMA Item 16b: Cite studies that appeared to meet inclusion criteria but were excluded, with reasons.]
### 3.2 Study Characteristics [PRISMA Item 17]
**Table: Characteristics of Included Studies**
| Study | Country | Design | Population (N) | Intervention | Comparator | Outcome(s) | Follow-up | Funding |
|-------|---------|--------|----------------|-------------|-----------|------------|-----------|---------|
| Author1 (Year) | [country] | [design] | [N] | [intervention] | [comparator] | [outcomes] | [duration] | [source] |
| Author2 (Year) | ... | ... | ... | ... | ... | ... | ... | ... |
### 3.3 Risk of Bias in Studies [PRISMA Item 18]
**Table: Risk of Bias Summary (Traffic-Light)**
| Study | D1 | D2 | D3 | D4 | D5 | Overall |
|-------|----|----|----|----|----|---------|
| Author1 (Year) | [L/S/H] | [L/S/H] | [L/S/H] | [L/S/H] | [L/S/H] | [L/S/H] |
| Author2 (Year) | ... | ... | ... | ... | ... | ... |
L = Low risk, S = Some concerns, H = High risk
[Narrative summary of risk of bias findings across studies]
### 3.4 Results of Individual Studies [PRISMA Item 19]
**Table: Individual Study Results**
| Study | Outcome | Effect Estimate | 95% CI | p-value | Weight |
|-------|---------|----------------|--------|---------|--------|
| Author1 (Year) | [outcome] | [estimate] | [lower, upper] | [p] | [%] |
| Author2 (Year) | ... | ... | ... | ... | ... |
### 3.5 Results of Syntheses [PRISMA Items 20a-20d]
#### Primary Outcome: [Name]
**20a. Study characteristics**: [Brief summary of contributing studies' characteristics and risk of bias]
**20b. Pooled result**:
- Pooled effect: [estimate] (95% CI: [lower, upper])
- Statistical significance: Z = [value], p = [value]
- Heterogeneity: I² = [value]% (95% CI: [lower, upper]), Q = [value] (df = [n], p = [value]), tau² = [value]
- Prediction interval: [lower, upper]
**Forest Plot**: [Insert or reference forest plot]
**20c. Heterogeneity investigation**:
- Subgroup analyses: [results]
- Meta-regression: [results, if conducted]
**20d. Sensitivity analyses**:
1. Leave-one-out: [results — did any single study substantially change the estimate?]
2. Excluding high-risk studies: [revised estimate]
3. Fixed vs. random effects: [comparison]
#### Secondary Outcome(s): [Name]
[Repeat structure above for each secondary outcome]
### 3.6 Reporting Biases [PRISMA Item 21]
[Report assessments of publication bias:]
- Funnel plot: [description of symmetry/asymmetry]
- Statistical test: [result]
- Trim-and-fill: [adjusted estimate, if applicable]
- Other assessments: [protocol-outcome comparison]
### 3.7 Certainty of Evidence [PRISMA Item 22]
**GRADE Summary of Findings Table**
| Outcome | Studies (n) | Participants (N) | Effect (95% CI) | Certainty | Rationale |
|---------|------------|-------------------|-----------------|-----------|-----------|
| [Outcome 1] | [n] | [N] | [estimate (CI)] | [High/Moderate/Low/Very Low] | [Reasons for up/downgrading] |
| [Outcome 2] | [n] | [N] | [estimate (CI)] | [level] | [reasons] |
---
## 4. Discussion [PRISMA Item 23]
### 4.1 Summary of Evidence
[Provide a general interpretation of the results in the context of other evidence. Address:]
- Main findings for each outcome
- How findings compare to previous reviews
- Consistency of findings across studies
### 4.2 Limitations
**Limitations of the evidence**:
- [e.g., risk of bias across studies, inconsistency, indirectness, imprecision]
**Limitations of the review process**:
- [e.g., language restrictions, database coverage, inability to contact authors]
### 4.3 Implications
**For practice**:
- [What practitioners should do based on these findings]
**For research**:
- [Gaps identified, recommended future study designs]
**For policy**:
- [Policy implications, if applicable]
---
## 5. Other Information
### 5.1 Registration and Protocol [PRISMA Item 24]
[Provide registration information and link to protocol:]
- Registry: [e.g., PROSPERO]
- Registration number: [number]
- Protocol URL: [link]
- Deviations from protocol: [describe any deviations and rationale]
### 5.2 Support [PRISMA Item 25]
[Describe sources of support:]
- Financial: [funding sources and grant numbers]
- Non-financial: [e.g., institutional support, access to databases]
- Role of funder: [describe any role of funders in the review]
### 5.3 Competing Interests [PRISMA Item 26]
[Declare competing interests of all authors]
### 5.4 Availability of Data and Materials [PRISMA Item 27]
[Report availability of:]
- [ ] Data extraction forms
- [ ] Extracted data from included studies
- [ ] Analysis code
- [ ] List of excluded studies with reasons
- [ ] PRISMA checklist (completed)
---
## Appendices
### Appendix A: Full Search Strategies
[Complete search strategies for all databases]
### Appendix B: Excluded Studies with Reasons
| Study | Reason for Exclusion |
|-------|---------------------|
| [citation] | [reason] |
### Appendix C: PRISMA 2020 Checklist
[Completed PRISMA 2020 checklist with page/section numbers for each item]
| Item # | Checklist Item | Reported on Page/Section |
|--------|---------------|-------------------------|
| 1 | Title | [page] |
| 2 | Abstract | [page] |
| ... | ... | ... |
| 27 | Availability | [page] |

93
skills/deep-research/templates/research_brief_template.md Normal file
View File

@@ -0,0 +1,93 @@
# Research Brief Template
## Purpose
Standard output format for `quick` mode research. Concise, actionable, evidence-based.
## Template
```markdown
# Research Brief
**Title**: [Descriptive title in sentence case]
**Date**: [YYYY-MM-DD]
**Mode**: Quick Research Brief
**AI Disclosure**: This brief was produced with AI-assisted research tools. All findings are verified against cited sources.
---
## Executive Summary
[100-150 words. State the question, key findings, and main implication.]
---
## Background & Research Question
### Context
[2-3 paragraphs providing necessary background for the reader.]
### Research Question
> [The refined research question, as a single clear sentence.]
### Scope
- **In scope**: [what this brief covers]
- **Out of scope**: [what this brief does not cover]
---
## Key Findings
### Finding 1: [Title]
[2-3 sentences with citation(s)]
**Evidence strength**: [Strong / Moderate / Emerging]
**Source(s)**: [Brief citation]
### Finding 2: [Title]
[2-3 sentences with citation(s)]
**Evidence strength**: [Strong / Moderate / Emerging]
**Source(s)**: [Brief citation]
### Finding 3: [Title]
[2-3 sentences with citation(s)]
**Evidence strength**: [Strong / Moderate / Emerging]
**Source(s)**: [Brief citation]
[Add more findings as needed, typically 3-5]
---
## Analysis & Implications
### What This Means
[2-3 paragraphs interpreting the findings. Connect evidence to practical implications.]
### Recommendations
1. [Actionable recommendation with evidence basis]
2. [Actionable recommendation with evidence basis]
3. [Actionable recommendation with evidence basis]
---
## Limitations
- [Limitation 1: e.g., search scope, time constraints]
- [Limitation 2: e.g., evidence gaps]
- [Limitation 3: e.g., AI-assisted analysis limitations]
---
## References
[APA 7.0 format, 5-10 references typical for quick mode]
1. [Reference 1]
2. [Reference 2]
...
```
## Usage Notes
- Target length: 500-1,500 words
- Minimum 5 cited sources
- Every finding must have at least one citation
- Evidence strength ratings should align with source quality hierarchy
- Keep language accessible — this is a brief, not a journal article