agentleFS
Google CloudKnown publisherGoogleCloudPlatform/devrel-demos/agents/newsroom-harness/SKILL.md

newsroom-harness

Skill397 starsChanged 10 days agoApache-2.0View on GitHub

Slip into the Newsroom agent harness — combines an investigative reporter persona (OCeaN), traffic-light permissions, graph memory with dreaming, local reporting skills, and three deterministic referees (Truth, Safety, and Reading grade).

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 2 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use this skill whenever the user asks you to "slip into the Newsroom harness", investigate a breaking news story, fact-check a draft, or …

    line 9
  2. Very specific:

    Read identity/persona.md, identity/job.md, identity/style.md, and permissions.md.

    line 14
  3. Very specific:

    Check memory/memory_graph.json (./tools/newsroom memory) so you never repeat debunked rumors or forget verified primary sources.

    line 16
  4. Very specific:

    Use skills/verify-sources/SKILL.md to triangulate facts.

    line 18
  5. Very specific:

    Use skills/write-bulletin/SKILL.md to write the bulletin.

    line 19
  6. Distinctive:

    Run the harness grader on your draft:

    line 21
  7. Very specific:

    If Truth Grade, Safety Grade, or Reading Grade is below 80/100 (or exits non-zero), fix every listed violation and re-run until all three …

    line 25
  8. Distinctive:

    After completing a story, run:

    line 27

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Newsroom Harness (newsroom-harness)

Overview

Use this skill whenever the user asks you to "slip into the Newsroom harness", investigate a breaking news story, fact-check a draft, or run the newsroom grading loop.

Mandatory Turn-by-Turn Harness Loop

  1. Step 1 — Equip Identity & Permissions: Read identity/persona.md, identity/job.md, identity/style.md, and permissions.md.
  2. Step 2 — Consult Memory Graph: Check memory/memory_graph.json (./tools/newsroom memory) so you never repeat debunked rumors or forget verified primary sources.
  3. Step 3 — Execute Local Skills:
    • Use skills/verify-sources/SKILL.md to triangulate facts.
    • Use skills/write-bulletin/SKILL.md to write the bulletin.
  4. Step 4 — Run the Referee Scoreboard (Hard Gate): Run the harness grader on your draft:
    ./tools/newsroom grade path/to/draft.md
    If Truth Grade, Safety Grade, or Reading Grade is below 80/100 (or exits non-zero), fix every listed violation and re-run until all three grades pass.
  5. Step 5 — Dream (Memory Consolidation): After completing a story, run:
    ./tools/newsroom dream
    to promote verified scratchpad facts into the permanent memory graph and discard unverified noise.

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.