agentsclimarketplace

Research

Skill TheWatcher01/skills/.claude/skills/research

Structured multi-source research protocol with data freshness validation — use when gathering factual data from the web, GitHub, npm, or docs to produce sourced, timestamped, cross-referenced findingsFrom its SKILL.md

Install
npx -y skills add TheWatcher01/skills --skill research

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.8 KB, 592 tokens by cl100k_base, as published. Nobody here has run it

Research Skill — chainskills

Protocol

Step 1 — Clarify scope

Define the research questions before searching. Map each question to a decision it informs.

Step 2 — Memory check

Query conversation history for prior research on overlapping topics. Reuse FRESH findings (< 90 days). Skip Step 4 for already-verified facts.

Step 3 — Workspace scan

Read: package.json deps, ROADMAP, AGENTS.md, relevant source files, existing templates. Note current state before looking externally.

Step 4 — External research (parallel where possible)

For each question, query >=2 independent sources from the authority hierarchy:

  1. Official registry (npm, crates.io, PyPI)
  2. Official docs / spec / RFC
  3. Official GitHub repo (README, releases, CHANGELOG)
  4. Verified community resource (MDN, caniuse, awesome-*)
  5. Tech blog / article
  6. Forum / StackOverflow

AI memory alone is never sufficient — always fetch from sources 1-4.

Step 5 — Cross-reference

For critical facts (versions, breaking changes, security), verify with >=2 independent sources. Flag discrepancies: DISCREPANCY: source A says X, source B says Y.

Step 6 — Freshness stamp

Assign a status to every external claim:

  • FRESH — retrieved < 90 days ago
  • AGING — 90 days to 1 year
  • STALE — 1 to 2 years
  • EXPIRED — > 2 years
  • UNVERIFIED — not fetched, AI memory only

Step 7 — Structured output

Produce a report with:

  • Source table (claim / source / URL / date / freshness / confidence H|M|L)
  • Dependency audit table (package / pinned / latest / gap / advisory)
  • Workspace vs external delta (what the codebase assumes vs what is actually true)
  • Stale/unverified warnings
  • Recommended next steps

Anti-patterns

  • Do not cite a URL without accessing it — always fetch
  • Do not treat AI training data as a source — always verify
  • Do not draw single-source conclusions for critical decisions
  • Do not omit retrieval dates
  • Do not start external research before checking workspace state

Output template

### Research Report — {topic} — {YYYY-MM-DD}

#### Workspace Findings
{What the codebase already knows/does about this topic}

#### External Findings
| Claim | Source | URL | Date | Freshness | Confidence |
|-------|--------|-----|------|-----------|------------|

#### Dependency Audit
| Package | Pinned | Latest | Gap | Advisory |

#### Stale / Unverified
{Items to re-verify before acting}

#### Recommended Next Steps
{Handoff-ready actions}

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 326,679. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.