agentleFS
jqaisystemsjqaisystems/jqai-ai-skills/skills/research-brief-curator/SKILL.md

research-brief-curator

Skill3 starsChanged 6 months agoMITView on GitHub

Curate public-source links, saved notes, newsletters, documentation updates, release notes, market signals, and research snippets into safe daily or weekly briefs with source tracking, public/private visibility labels, archive decisions, and human review. Use when the user asks to make a research brief, summarize saved links, build an AI news digest, organize reading notes, create a public-safe source roundup, prepare a weekly intelligence brief, or separate public-ready notes from private research without exposing credentials, raw prompts, logs, local paths, account data, client material, or private source lists.

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 7 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill to turn public links and approved notes into a short, reviewable research brief.

    line 10
  2. Very specific:

    Read references/public-source-boundaries.md before using user-provided links, notes, exports, screenshots, or copied text.

    line 15
  3. Very specific:

    Read references/brief-template.md before drafting the final daily or weekly brief.

    line 17
  4. Very specific:

    Use scripts/build_brief_skeleton.py when the user provides structured JSON records or wants a repeatable Markdown skeleton.

    line 28
  5. Distinctive:

    Use public links, user-approved notes, and fake examples only.

    line 38
  6. Distinctive:

    Do not expose API keys, tokens, cookies, browser profiles, raw prompts, private source lists, logs, account data, dashboards, exports, …

    line 39
  7. Distinctive:

    Do not summarize paywalled, private, login-only, confidential, or internal-only material for public output unless the user explicitly …

    line 40
  8. Very specific:

    Do not present uncertain claims as facts.

    line 41
  9. Distinctive:

    Do not auto-post, auto-send, create public tickets, update a CMS, or push changes.

    line 42
  10. Distinctive:

    Keep quotes short and attribute them.

    line 43
  11. Distinctive:

    If the user provides risky material, do not include it in the brief.

    line 57

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Research Brief Curator

Overview

Use this skill to turn public links and approved notes into a short, reviewable research brief. Keep the source trail visible, label each item by visibility, and require human approval before anything is posted, sent, archived publicly, or reused as client-facing material.

Workflow

  1. Clarify the brief scope: topic, audience, time window, output format, and whether the brief is internal-only, public-ready, client-facing, or draft research.
  2. Read references/public-source-boundaries.md before using user-provided links, notes, exports, screenshots, or copied text.
  3. Normalize every item into a simple record: title, source, URL or citation, date, topic, visibility, status, summary, why it matters, and recommended next action.
  4. Read references/brief-template.md before drafting the final daily or weekly brief.
  5. Separate records into:
    • public_ready: safe to summarize with source link after review.
    • needs_review: claims, wording, or source quality need checking.
    • private_research: useful internally but not suitable for public reuse.
    • archive: duplicate, stale, or background material.
  6. Draft the brief with source-aware language and short decision notes. Do not copy long excerpts from sources.
  7. End with a safety verdict: BLOCK, REVIEW, or READY, plus the remaining manual checks.

Optional Script

Use scripts/build_brief_skeleton.py when the user provides structured JSON records or wants a repeatable Markdown skeleton.

python scripts/build_brief_skeleton.py items.json --title "Weekly AI Research Brief" --date 2026-05-28 --output brief.md

The script does not fetch URLs or publish anything. It only formats provided records and flags suspicious local paths or credential-like terms for review.

Safety Rules

  • Use public links, user-approved notes, and fake examples only.
  • Do not expose API keys, tokens, cookies, browser profiles, raw prompts, private source lists, logs, account data, dashboards, exports, local paths, client names, or private implementation details.
  • Do not summarize paywalled, private, login-only, confidential, or internal-only material for public output unless the user explicitly confirms it is approved.
  • Do not present uncertain claims as facts. Mark them needs_review and ask for source verification.
  • Do not auto-post, auto-send, create public tickets, update a CMS, or push changes. Return a brief and review checklist.
  • Keep quotes short and attribute them. Prefer paraphrase and links over copied text.

Output Format

Return:

  1. Safety verdict: BLOCK, REVIEW, or READY.
  2. Brief title, date, audience, and source count.
  3. Public-ready highlights with source links.
  4. Needs-review items and why they are held.
  5. Private-research or archive decisions.
  6. Recommended next actions.
  7. Manual review checklist before publishing or sharing.

If the user provides risky material, do not include it in the brief. Explain the minimum safe rewrite or redaction needed.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.