Zsun79/LitReviewSkill / skills/write-literature-review
write-literature-review
VerifiedIterative literature-review workflow for a research topic. Use this skill to draft up to 10 search keywords, build a seed set from OpenAlex title-and-abstract matches, expand by backward and forward citations, screen candidates by title and abstract, repeat until no new relevant papers remain, rank the final set, inspect up to 30 full texts, build a concept-reference knowledge graph, and write a literature review grounded only in the final included references.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 5.5 KB |
| agents/openai.yaml | config | 238 B |
| references/citation format/APA.md | markdown | 846 B |
| references/citation format/Chicago Author-Date.md | markdown | 903 B |
| references/citation format/Harvard.md | markdown | 938 B |
| references/citation format/IEEE.md | markdown | 808 B |
| references/citation format/MLA.md | markdown | 723 B |
| references/citation format/Vancouver.md | markdown | 748 B |
| references/review_template.md | markdown | 2.2 KB |
| references/review_writing.md | markdown | 4.0 KB |
| references/workflow.md | markdown | 8.7 KB |
| references/workflow_logging.md | markdown | 3.2 KB |
| scripts/build_knowledge_graph.py | script | 7.7 KB |
| scripts/lit_review_pipeline.py | script | 30.0 KB |
references/workflow_logging.md
View on GitHub# Workflow Logging Guide
## Purpose
Use `logs/workflow_log.md` as the running audit trail for the full literature-review workflow.
The log should make it possible for another reviewer to answer:
- what the agent did
- in what order it happened
- what files or commands were used
- what each step produced
- what was learned
- why the next decision was made
## Required File
Write the log to:
- `logs/workflow_log.md`
Create the file near the beginning of the workflow and keep appending to it as work happens.
## Logging Rules
- Log every major workflow step.
- Log every expansion or screening round separately.
- Record counts whenever a step changes the candidate set.
- Record failures, retries, unavailable full text, and scope changes explicitly.
- Keep the log factual and auditable.
- Do not replace the log with a polished summary only. It should remain step-by-step.
## Minimum Fields Per Entry
Each entry should include:
- step number or stage name
- timestamp or round identifier when available
- action taken
- inputs used
- output files updated
- key result or counts
- decision or next action
## Suggested Template
```md
# Workflow Log
## Step 1 - Scope Setup
- Action: clarified review scope and created `plan/review_plan.md`
- Inputs: user topic, inclusion criteria, time window
- Outputs: `plan/review_plan.md`
- Result: finalized scope for the review
- Next: draft keyword list
## Step 2 - Keyword Drafting
- Action: drafted search keywords
- Inputs: topic statement and synonyms
- Outputs: `search/seed_keywords.json`
- Result: selected 8 keywords with high recall focus
- Notes: dropped 2 overly broad keywords because they caused topic drift
- Next: run seed search
## Step 3 - Seed Search
- Action: searched OpenAlex title and abstract fields
- Inputs: keyword set from `search/seed_keywords.json`
- Outputs: `search/seed_candidates.jsonl`, `search/candidates.jsonl`
- Result: 184 seed records retrieved, 161 unique after dedupe
- Next: screen seed set
## Step 4 - Screening Round 1
- Action: screened title and abstract of seed set
- Inputs: `search/seed_candidates.jsonl`
- Outputs: `screening/filtered_candidates.jsonl`, `screening/visited_titles.jsonl`
- Result: 161 screened, 54 included, 107 excluded
- Notes: common exclusion reasons were wrong domain and non-scholarly publication type
- Next: expand from filtered set
## Step 5 - Expansion Round 1
- Action: expanded by backward and forward citations
- Inputs: `screening/filtered_candidates.jsonl`
- Outputs: `expansion/expanded_candidates.jsonl`, `screening/screening_queue.jsonl`
- Result: 420 expanded records, 233 new unvisited candidates after dedupe
- Next: screen expansion queue
```
## End-of-Workflow Summary
End the log with a short final section containing:
- total number of search or expansion rounds
- total number of unique screened candidates
- final included reference count
- number of full texts actually read
- main themes discovered
- key limitations or blind spots
## Relationship to Final Outputs
The log is not the literature review itself.
Keep these separate:
- `logs/workflow_log.md`: operational audit trail
- `review/literature_review.md`: editable review draft
- `literature_review.pdf`: final root deliverable
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |