Index of every page on this site: llms.txt
All skills
vinta avatar

/preview-verdicts

@4c3cb15
by Vinta Chenvinta/awesome-python336k stars
29,136
Run it yourself

Generate the interactive keep/remove/add/reject verdict preview (HTML page with per-row feedback controls) whenever a prune sweep, batch entry edit, or restructure needs maintainer review before touching README.md — and process the feedback JSON the maintainer pastes back.

Use this Skill: https://skilld.dev/gh/vinta/awesome-python/preview-verdicts

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈73 tokens for metadata: the name and description. ≈942 when used: this file.

Description uses 3.6% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈292 characters, or 3.6%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Verdict preview

Maintainer review happens through an interactive HTML page: one row per entry with your seeded verdict and reason, verdict buttons and a reason field for the maintainer, and a Copy feedback button that exports only changed or commented rows as JSON. Generate the page, wait for the pasted JSON, then apply it. Entry changes land in README.md only after the review — and only on an explicit go.

Generate the preview

  1. Build the DATA array. A group is [section, subcategory, rows]; a row is [entry, url, downloads, kind, verdict, reason], plus an optional seventh source field.
    • source says where a row came from when that is not obvious, such as the starred repo behind a scan candidate or an incumbent pulled in during the audit. Name each repo as owner/repo (starred: piskvorky/smart_open), which the page turns into a GitHub link. The page shows a Source column only when a row has one.
    • subcategory may carry a note after — (rendered muted): use it for proposed splits, re-homes, or anything the maintainer should weigh for the whole group.
    • downloads is PyPI last-month as a comma-formatted string; use — when no signal exists (e.g. agent skill packs), stdlib for standard-library modules, fetch failed when the lookup failed. State the fetch date in the sub-header.
    • kind sets the buttons independently of the seeded verdict:
      • listed: keep, remove (taken out of README.md), or park.
      • candidate: add, reject (not added, logged under "Rejected Additions"), or park.
      • structure: restructure (take the change the reason or group note describes), park, or reject (logged under "Rejected Additions").
    • Seed verdict from the current adjudication or dry-run. park defers a decision for a later review; preserve the row's kind so its original choices remain available.
    • Show a replacement as two rows: the old entry listed / remove, the new entry candidate / add.
    • reason is plain language the maintainer reads cold — no invented shorthand. When fresh evidence contradicts the seeded verdict (a big download count on a remove or reject, a dead repo on a keep), say so in that row's reason instead of silently changing the seed.
  2. Copy template.html (sibling of this file) and replace the placeholders: __TITLE__ (page title), __SUB__ (sub-header: scope, seed provenance, fetch date, and the standing instruction to flip/comment then Copy feedback), __KEY__ (localStorage key), __DATA__ (the array). __KEY__ must be unique per review — slug plus date, e.g. awesome-python-science-2026-09-01 — because saved state under a reused key bleeds a previous review's flips into rows with the same section and entry name.
  3. Write the page to local/awesome-python-<slug>-preview.html in the repo root (git-ignored; create the directory if needed), open it, and tell the maintainer the path and the return path: flip or comment rows (they highlight yellow), press Copy feedback, paste the JSON into the chat. Done when the page is open and the return path is stated.

Process the pasted feedback

Each JSON row is {section, subcategory, entry, kind, my_verdict, your_verdict, reason}, with kinds and verdicts as defined above. The maintainer's verdict is final — apply it, never re-argue it. An empty reason means the verdict stands unexplained; that is enough.

Before executing, surface anything the flips imply that the maintainer has not decided: a use case pushed past its cap, an entry left homeless by a proposed split, a request that is already satisfied (a no-op). Ask, then execute on their go. Done when every pasted row is either applied or surfaced back — none silently dropped.

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

1 warning28d3 checks · Risk SAFE
  • Gen Agent Trust Hub28d

    The skill generates an interactive local HTML page to review changes before applying them to a repository. It presents a potential security risk as it uses unsanitized data from the README being audited to construct this page, which could lead to Cross-Site Scripting (XSS) in the maintainer's browser.

  • Socket1mo

    No alerts

  • Snyk1mo

    Risk: MEDIUM · 1 issue

Signed by skilld at 4c3cb15. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 2 weeks ago

README badge

README badge for vinta/awesome-python/preview-verdicts