marciob/skill-research-papers / research-papers
research-papers
VerifiedResearch, filter, compare, and synthesize scientific papers across arXiv, OpenAlex, ACM Digital Library, DBLP, Semantic Scholar, PubMed, SSRN, and publisher landing pages. Use when the agent needs to do literature discovery, find top or recent papers, build a reading list, compare methods or results, trace citation graphs, summarize evidence, or produce properly attributed paper recommendations and literature reviews.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 14.5 KB |
| agents/openai.yaml | config | 221 B |
| references/api-playbook.md | markdown | 16.9 KB |
| references/output-shapes.md | markdown | 2.5 KB |
| references/screening-rubric.md | markdown | 3.0 KB |
| references/smoke-tests.md | markdown | 4.4 KB |
| references/source-guides.md | markdown | 4.9 KB |
| scripts/.gitignore | other | 7 B |
| scripts/fetch_and_parseexecutable | script | 990 B |
| scripts/fetch_and_parse.pyexecutable | script | 52.0 KB |
| scripts/requirements.txt | manifest | 496 B |
| scripts/smoke-test-sources.shexecutable | script | 5.8 KB |
references/output-shapes.md
View on GitHub# Output Shapes
Use this file when deciding how to present results to the user.
## Quick Shortlist
Use when the user asks for "top papers," "best papers," or a reading list.
Template:
- Scope: one sentence on topic and time window
- Best picks: 5 to 8 papers
- For each paper: citation, link, one-line reason for inclusion, status marker such as published, survey, or preprint, evidence depth tag (`[Full text]`, `[Abstract only]`, or `[Metadata only]`), and full-text URL when available
- Closing note: what was excluded and why
## Comparative Review
Use when the user asks to compare methods, schools of thought, or historical progression.
Template:
- Question answered
- Search coverage
- Comparison table or structured bullets — include evidence depth tag per paper and full-text URL when available
- Key differences: methods, data, assumptions, findings, limitations
- Distinguish claims sourced from full text versus abstract-only papers
- Recommendation: which papers to read first and why
## Evidence Brief
Use when the user wants to know what the literature says about a claim.
Template:
- Bottom line first
- Strength of evidence — note how many papers were assessed from full text vs abstract only
- Strongest supporting papers — include evidence depth tag and full-text URL when available
- Conflicting or limiting papers — include evidence depth tag
- Distinguish claims backed by full-text reading from those inferred from abstracts
- Gaps and uncertainty
## Literature Map
Use when the field is broad or the user is new to it.
Template:
- Surveys and review papers — include evidence depth tag and full-text URL when available
- Seminal anchors — include evidence depth tag and full-text URL when available
- Strong recent papers — include evidence depth tag and full-text URL when available
- Benchmarks or datasets
- Open problems or disagreements
## Citation Hygiene
Always provide enough metadata for retrieval:
- Title
- Authors
- Year
- Venue or status
- Stable link
Flag these explicitly when relevant:
- Preprint
- Review or survey
- Benchmark or dataset paper
- Working paper
- Metadata-only assessment
## Evidence Depth Tags
Tag every paper in the output with one of:
- `[Full text]` — full paper was read and claims are sourced from the content
- `[Abstract only]` — only the abstract was available; claims are limited to what the abstract states
- `[Metadata only]` — only title, venue, and citation metadata were available
Include the full-text URL alongside the paper link when available, so the user can access the same source.
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 |