Grades are a hygiene check, not a security guarantee.

Claude Code, Skill

skills

My personal collection of AI agent skills for day-to-day development, installable and managed with skills.sh.

Compatibility: Not stated.

Grade C

One Medium finding. Read the report before installing.

Graded by PlugVetRubric v2

Not a guarantee. A hygiene check of the source at the commit shown, not a security audit.

AGENT_IGNORE_INSTRUCTIONS

Why this grade

  • AGENT_IGNORE_INSTRUCTIONS Medium

    Text tells the agent to ignore earlier instructions.

    docs/plans/0017-plan-promptkit-2026-08-07.md, line 15 **Success:** paste a half-formed instruction, get back a prompt you can send without editing, plus two lines telling you what it resolved against the real repo and what it assumed

Any High, or two or more Medium findings, give D. One Medium gives C. Low findings or a security Note give B. Each code names a finding behind the letter.

Grades are a hygiene check of the source at the commit shown, on the date checked, not a security audit or a guarantee. The free directory grades are automated. Before release, a PlugVet engineer reviews every paid report and any grade we hold back, including all F grades, against the source code and the evidence we quote.

What was checked

Plain notes from the hygiene checks.

  • Permissions

    Pass

    No shell or command call was found in the code. This check does not read permissions an entry declares. A permission line to run shell commands, write or edit files and fetch web content was found in this repository, and it may apply to this entry.

  • Install and update

    Pass

    No install-time script found, and dependencies are pinned or absent.

  • Network

    Concern

    can reach a host that is not written as a fixed address. can reach rdap.org. The description does not name that host. can reach registry.npmjs.org. The description does not name that host.

    NETWORK_NO_FIXED_HOST NETWORK_UNNAMED_HOST

    scripts/security.sh, line 108

    if shits "$body" '\brm -rf\b|\bsudo \b|\bchmod 777\b|:\(\)\{|\beval \$|curl [^|]*\| *(sh|bash)|wget [^|]*\| *(sh|bash)'; then

    skills/namekit/SKILL.md, line 34

    curl -s -o /dev/null -w '%{http_code}' -L --max-time 10 https://rdap.org/domain/<name>.com

    skills/namekit/SKILL.md, line 37

    curl -s -o /dev/null -w '%{http_code}' --max-time 10 https://registry.npmjs.org/<name>

    Show more

    skills/seokit/bin/extract.mjs, line 61

    const res = await fetch(url, {

  • Secrets

    Concern

    reads a credential or environment file.

    SECRETS_PATH_DOTENV

    scripts/security.sh, line 103

    if shits "$body" '\b(read|cat|print|log|echo|exfiltrate|collect)\b[^.`]{0,40}\b(secret|token|api[_ -]?key|password|credential|\.env|private key)'; then

    skills/wikikit/SKILL.md, line 78

    documents: [package.json, src/index.ts, .env.example]

    docs/plans/0024-plan-testkit-2026-08-18.md, line 42

    | Integration-test environment | **Refuse any target testkit has not verified as disposable, then offer to scaffold one.** A test may only run against localhost, a container testki

    Show more

    docs/plans/0024-plan-testkit-2026-08-18.md, line 90

    - **Verify any datastore or service target as disposable before running against it**, and refuse otherwise. Localhost, a container, or an explicit test URL qualifies; an ambient `.

    docs/plans/0024-plan-testkit-2026-08-18.md, line 108

    - [ ] A target named only by an ambient `.env` value is treated as unverified.

    docs/plans/0035-plan-seokit-2026-10-04.md, line 49

    | Production origin | `local` runs need the production origin to judge canonical, `og:url`, hreflang, and sitemap URLs. seokit reads it from the repository first: framework config

    docs/wiki/reference.md, line 139

    | Secret handling | Snyk | med | "read/print/log … secret/token/API key/password/`.env`" |

    skills/handoffkit/SKILL.md, line 26

    - **Redact**: API keys, tokens, passwords, and personally identifiable information. Never carry secrets into the document. Refer to them by name ("the staging DB password in `.env`

  • Agent instructions

    Serious

    instructions tell the agent to ignore earlier instructions.

    AGENT_IGNORE_INSTRUCTIONS

    docs/plans/0017-plan-promptkit-2026-08-07.md, line 15

    **Success:** paste a half-formed instruction, get back a prompt you can send without editing, plus two lines telling you what it resolved against the real repo and what it assumed

  • Obfuscation

    Pass

    No encoded or minified payload found.

  • Licence

    Pass

    A licence file is present (LICENSE).

  • Maintainer activity

    Pass

    The latest commit is within the last 365 days.

Get the full report on this tool, £49

Note

Hygiene check of commit c70734b27755. can reach a host that is not written as a fixed address. can reach rdap.org. The description does not name that host. can reach registry.npmjs.org. The description does not name that host. reads a credential or environment file. instructions tell the agent to ignore earlier instructions.

Date and source

Date checked:

View source

github.com

Commit graded: c70734b27755658c2bf38f3a4bb073f4248b6cbf

This link goes to the original host. PlugVet does not host or check what you download from there, and installing it is your own decision and risk.

Read how this letter is chosen