agentleFS
yiyaw-labyiyaw-lab/agent-armor/publish/SKILL.md

publish

Skill1 starsChanged 5 months agoMITView on GitHub

Schema-first publisher for yiya.dev. Turn one durable, public-safe artifact from a Claude Code session into a typed content entry, after disclosure checks, a secrets scan, and the four-part quality gate. Reads live yiyaw-lab/yiyadev schemas, proves the entry with npm run build, then opens a PR or compare URL. Never merges. Modes: <artifact|path> | sweep | draft.

  • Reads credentials
  • Installs packages

60 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use create, render and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • MicrosoftSkill

    create

    Also in 4 projects that use this

  • suxrobGMSkill

    render

    Generate raster images (PNG, JPEG, WebP) from natural language prompts using Python rendering engines. Covers graphics, charts, layouts, and procedural art. Not for SVG output.

    Also in 3 projects that use this

  • JaimeGBASkill

    ingest

    Digest an external source (article, paper, transcript, URL, book chapter) and integrate its content into the knowledge base

    Also in 6 projects that use this

  • genomewalkerSkill

    explore

    RLM-style recursive memory exploration - dynamically navigate the memory graph

    Also in 4 projects that use this

  • GoogleSkill

    translate

    Translates a common phrase into one or more languages.

    Also in 3 projects that use this

  • VercelSkill

    setup

    Set up Vercel CLI and project configuration. Use when the user says "set up Vercel", "configure Vercel", "link to Vercel", or "vercel init".

    Also in 14 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. 0 are common, 5 are distinctive, and 9 are very specific to this project.

  • 0

    Common

    Most of the 49 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 9

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Distinctive:

    Run disclosure and secrets checks before rendering anything public.

    line 11
  2. Very specific:

    Read the live yiyaw-lab/yiyadev content files before writing.

    line 14
  3. Very specific:

    Use an existing checkout first.

    line 20
  4. Distinctive:

    If the private clone or push fails because credentials are unavailable, stop with the exact files/patch prepared and the GitHub compare URL …

    line 38
  5. Very specific:

    Always re-read src/content/*.ts.

    line 42
  6. Distinctive:

    Use for reusable lessons and hard-won operating principles.

    line 49
  7. Very specific:

    Use only for public repos with real receipts in href and tags.

    line 56
  8. Very specific:

    Use only for small public builder tools.

    line 63
  9. Very specific:

    Use for substantial notes or essays.

    line 87
  10. Very specific:

    Use for dated bets and scored outcomes.

    line 106
  11. Very specific:

    Run the disclosure boundary and secrets scan.

    line 111
  12. Very specific:

    Create a branch in the yiyadev checkout:

    line 119
  13. Distinctive:

    If no candidate clears the bar, say so and open no PR.

    line 160
  14. Distinctive:

    Run resolve, disclosure scan, gate, classification, live schema read, and rendering.

    line 164

“Common” is measured: kept by at least 6 in 10 of the 49 edited versions of this file we could read. “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

Publish is the public-facing sibling of the harvest -> milk -> vault flywheel. It promotes only the few artifacts that are safe, specific, reusable, and verified enough to belong on yiya.dev. It never invents work and never merges its own PR.

Non-Negotiables

  • Run disclosure and secrets checks before rendering anything public.
  • Publish receipts, not strategy. Redact or reject private project details, internal IDs, credentials, customer/user data, unshipped plans, and private taxonomy.
  • Enforce the four-part gate: Reusable, Non-trivial, Specific, Verified. Reject with named reasons when any part fails.
  • Read the live yiyaw-lab/yiyadev content files before writing. Treat every schema shown below as a snapshot, not authority.
  • Prove every content edit with npm run build before opening a PR.
  • Open a PR or provide a compare URL. Never push to main, merge, enable auto-merge, or batch unrelated artifacts.

Target Repo Setup

Use an existing checkout first. Prefer YIYADEV_REPO, then ~/code/yiyadev, then ~/yiyadev, then /tmp/yiyadev.

target="${YIYADEV_REPO:-$HOME/code/yiyadev}"
if [ ! -d "$target/.git" ] && [ -d "$HOME/yiyadev/.git" ]; then target="$HOME/yiyadev"; fi
if [ ! -d "$target/.git" ] && [ -d "/tmp/yiyadev/.git" ]; then target="/tmp/yiyadev"; fi
if [ ! -d "$target/.git" ]; then
  if command -v gh >/dev/null 2>&1; then
    gh repo clone yiyaw-lab/yiyadev "$target"
  else
    git clone https://github.com/yiyaw-lab/yiyadev.git "$target"
  fi
fi
git -C "$target" fetch origin main
git -C "$target" checkout main
git -C "$target" pull --ff-only

If the private clone or push fails because credentials are unavailable, stop with the exact files/patch prepared and the GitHub compare URL that should be opened after authentication. Do not claim a PR exists unless it does.

Live Schema Snapshot

Always re-read src/content/*.ts; this snapshot was checked on 2026-07-02 and may drift.

  • src/content/guides.ts
    export type Guide = { title: string; body: string[] };
    export const guides: Guide[] = [];
    Use for reusable lessons and hard-won operating principles. Body is tight editorial prose, usually 2-4 paragraphs.
  • src/content/code.ts
    export type CodeRepo = { title: string; note: string; href: string; tags: string[] };
    export const code: CodeRepo[] = [];
    Use only for public repos with real receipts in href and tags.
  • src/content/tools.ts
    export type Tool = { name: string; note: string; get: string; href: string; tags: string[] };
    export const tools: Tool[] = [];
    Use only for small public builder tools. get is the human install/use line.
  • src/content/writing.ts
    export type WritingCategory =
      | "Prior" | "Prediction" | "Pattern" | "Rethought"
      | "Replicated" | "Remains" | "Recursion" | "Unfalsifiable";
    export type Essay = {
      slug: string;
      title: string;
      dek: string;
      category: WritingCategory;
      date: string;
      readTime: string;
      href?: string;
      origin?: string;
      syndicated?: { platform: string; url: string }[];
      body?: string;
      tags?: string[];
      shelfLife?: "perishable" | "perennial";
      draft?: boolean;
    };
    export const essays: Essay[] = [];
    Use for substantial notes or essays. Prefer an on-site body; use href/origin only when the canonical piece lives elsewhere.
  • src/content/track-record.ts
    export type TrackRecordStatus = "open" | "hit" | "partial" | "miss";
    export type TrackRecordEntry = {
      id: string;
      stated: string;
      claim: string;
      domain: "research" | "engineering" | "product";
      status: TrackRecordStatus;
      resolution?: string;
      resolved?: string;
      source?: string;
      sourceLabel?: string;
      draft?: boolean;
    };
    export const trackRecord: TrackRecordEntry[] = [];
    Use for dated bets and scored outcomes. open is allowed only for a real dated prediction. hit, partial, and miss require resolution, resolved, and evidence.

Mode: <artifact|path>

  1. Resolve one artifact. It can be a harvested finding, decision, session summary, build-log entry, scored prediction, shipped public repo, or explicit path. Read it fully.
  2. Run the disclosure boundary and secrets scan. Search for API keys, tokens, .env content, credentials, private absolute paths, emails, customer/user data, private strategy, internal taxonomy, and unshipped plans.
  3. Apply the four-part gate:
    • Reusable: useful beyond this one session.
    • Non-trivial: not generic advice or filler.
    • Specific: includes a concrete pattern, number, date, commit, test, repo, or outcome.
    • Verified: traceable to a real receipt.
  4. Classify the artifact to exactly one target file: guides, code, tools, writing, or track-record.
  5. Re-read the target file and nearby entries. Match its object shape, ordering, imports, trailing commas, and voice.
  6. Create a branch in the yiyadev checkout:
    slug="<short-kebab-slug>"
    git -C "$target" checkout -B "publish/$slug"
  7. Insert one typed entry. Keep the PR one artifact wide.
  8. Run:
    cd "$target"
    npm ci --silent 2>/dev/null || npm install --silent
    npm run build
    Fix malformed entries until the build is green. Never open a red PR.
  9. Commit and push:
    git -C "$target" add src/content
    git -C "$target" commit -m "Publish <short title>"
    git -C "$target" push -u origin "publish/$slug"
  10. Open review without merging:
if command -v gh >/dev/null 2>&1; then
  gh -R yiyaw-lab/yiyadev pr create --base main --head "publish/$slug" \
    --title "Publish <short title>" \
    --body "<source artifact, gate verdicts, validation command, and do-not-merge reminder>"
else
  echo "Open PR: https://github.com/yiyaw-lab/yiyadev/compare/main...publish/$slug?expand=1"
fi

Mode: sweep

Scan the newest harvest outputs and propose candidates. Do not write files until the user picks one.

For each candidate, report:

  • source artifact and receipt
  • disclosure/secrets verdict
  • four-part gate verdict
  • target content file
  • rendered draft object

If no candidate clears the bar, say so and open no PR. One excellent item beats a batch of weak ones.

Mode: draft <artifact|path>

Run resolve, disclosure scan, gate, classification, live schema read, and rendering. Print the exact typed object and target file. Do not commit, push, or open a PR. If useful, temporarily insert and run npm run build, then revert the local draft before stopping.

Voice

Match yiya.dev: austere, editorial, claim-free, receipts over adjectives. No hype, no vanity metrics, no generic AI boilerplate. If the artifact cannot be written without overclaiming, reject it.

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.