agentleFS
strelov1strelov1/freehire/.agents/skills/review-dictionary-requests/SKILL.md

review-dictionary-requests

Skill3.6k starsChanged yesterdayMITView on GitHub

Review Freehire dictionary requests from /dictionary, classify duplicates and ambiguous terms, implement useful vocabulary additions, and reconcile request statuses when authorized.

Adopted by

Steady: 0 new projects took it in the last 4 weeks, none the 4 weeks before.thin data

Week of Sep 20: 0 new projectsWeek of Sep 27: 0 new projectsWeek of Oct 4: 0 new projectsWeek of Oct 11: 0 new projectsWeek of Oct 18: 0 new projectsWeek of Oct 25: 0 new projectsWeek of Nov 1: 0 new projectsWeek of Nov 8: 0 new projects
Sep 2026Oct 2026Nov 2026
New projects taking a version of this file each week, over the 8 weeks our latest run read: only 0 in all, so this is a hint, not a trend.

Who's talking about strelov1/freehire

The biggest 5 of 35 mentions linking to the repository that our crawl found. Each line is their own headline.

See all 35

Talked about on GitHub and Bluesky (35 times).

On GitHub · 34

29 more on GitHub

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 6 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

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

    Read root AGENTS.md for current production access, .out-of-scope/README.md, and the AGENTS.md of the requested dictionary.

    line 13
  2. Distinctive:

    Use the established SSH/Postgres access.

    line 17
  3. Distinctive:

    Add alias: existing concept, new spelling.

    line 27
  4. Distinctive:

    Add concept: a distinct, useful skill, specialization or company industry.

    line 28
  5. Distinctive:

    Search dictionary mappings, descriptions, ambiguity gates, tests and exclusion comments.

    line 32
  6. Very specific:

    Use the existing thematic dictionary tables.

    line 46
  7. Distinctive:

    Test canonicalization and realistic prose for the accepted spellings.

    line 59
  8. Distinctive:

    When queue updates are authorized, close existing supported terms as accepted, and clear confirmed test noise or unsupported requests as …

    line 72
  9. Very specific:

    Use a transaction and explicit reviewed ids with WHERE status = 'pending' and RETURNING.

    line 75

≈ “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

Review dictionary requests

Work from the repository root. The public /dictionary page describes the vocabulary; submitted requests live in production dictionary_gaps, not on that page.

Inspect the queue

Read root AGENTS.md for current production access, .out-of-scope/README.md, and the AGENTS.md of the requested dictionary. Inspect the schema in migrations/0188_dictionary_gaps.sql, later migrations affecting it, and internal/platform/db/queries/dictionary_gaps.sql before querying. Use the established SSH/Postgres access; keep credentials out of output.

Fetch every pending row, including id, kind, raw_term, normalized_term, request_count, first_requested_at, last_requested_at and resolution_note. Fetch status totals separately. Dates are request dates, not proof of when a human last reviewed the queue. Avoid exposing submitter identities.

Complete the inspection with one disposition for every pending row:

  • Already supported: identify the canonical and verify actual canonicalization.
  • Add alias: existing concept, new spelling; preserve the canonical slug.
  • Add concept: a distinct, useful skill, specialization or company industry.
  • Needs clarification: multiple meanings, combined concepts, or insufficient evidence.
  • Reject: test noise, wrong vocabulary kind, or a documented exclusion that still applies.

Search dictionary mappings, descriptions, ambiguity gates, tests and exclusion comments. A string match alone does not prove a term resolves. Preserve the separation between skill, role, seniority and company industry. Broad occupational skills can be valid; being nontechnical alone is not grounds for rejection. Consult primary product documentation when a product's identity or scope is uncertain.

Implement authorized additions

A request to inspect or discuss produces a proposed disposition. A request to add useful terms authorizes the corresponding repository edits. Keep uncertain items pending and explain the specific ambiguity rather than inventing aliases.

For skills, follow internal/dict/skilltag/AGENTS.md: alias mappings, display labels where mechanical title casing is insufficient, and descriptions belong together. Use the existing thematic dictionary tables. After adding canonical skills, read internal/dict/skillvec/AGENTS.md and run go generate ./internal/dict/skillvec/ to append their permanent vector positions; preserve existing indices. Run make gen-contracts and commit the generated web vocabulary and descriptions too. Read the corresponding owners for specializations and industries instead of treating them as skills.

Choose canonical spellings deliberately. Correct a submitted typo into its actual concept; a one-off typo need not become a permanent prose alias. Check whether an alias is a common verb or ordinary phrase: design systems also occurs in "design systems that scale". Respect intentional exclusions and the distinction between structured-list canonicalization and prose parsing. New strong phrase matches can also corroborate nearby ambiguous skills; inspect that effect.

Test canonicalization and realistic prose for the accepted spellings. Add negative examples when ambiguity warrants them. Run the affected dictionary tests and the repository's applicable checks, including Go vet and integration-tag compilation. Review only this task's diff; preserve unrelated work.

Reconcile only what is live

Code prepared locally is not a deployed dictionary. Keep new requests pending until production resolves them. Deployment and catalogue backfill are separate actions: follow the user's authorized scope and the worker/deployment instructions before running either. Existing job facets need backfill and search refresh as documented by the dictionary owner.

When queue updates are authorized, close existing supported terms as accepted, and clear confirmed test noise or unsupported requests as rejected with an English resolution_note. For new additions, verify live resolution before acceptance. Use a transaction and explicit reviewed ids with WHERE status = 'pending' and RETURNING; compare changed rows with the intended set before committing. Preserve request counts and timestamps. Re-read the affected rows after committing. Accepted rows remain live under a partial unique index; rejected terms may be submitted again. Do not delete request history or bulk-close the pending queue.

Report added concepts and aliases separately, closed requests, items still pending, checks actually completed, and whether the result is local or deployed. Link the changed artifacts. Do not claim acceptance, deployment or backfill from local tests.

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.