Back to skill
Skillv2.3.6
ClawScan security
Test Publish Check · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMar 17, 2026, 6:57 PM
- Verdict
- benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- The skill's files and runtime instructions are consistent with a pre-publish checklist utility: it generates checklists and provides a small local CLI that stores logs/data under a user data directory; it does not request credentials or perform network exfiltration.
- Guidance
- This skill appears coherent and implements what it describes: checklist templates and a tiny local helper CLI. Notes before installing: (1) It will create a data directory (default ~/.local/share/test-publish-check) and append logs/data there — if you prefer, set TEST_PUBLISH_CHECK_DIR to a safe folder or run in a sandbox. (2) There are no network calls or credential requirements in the shipped scripts, but the registry/source listed in metadata is 'unknown' while SKILL.md references bytesagain.com/GitHub — verify the publisher if provenance matters. (3) If you want extra caution, inspect the two scripts yourself or run them in a disposable environment; otherwise this skill is internally consistent with its stated purpose.
Review Dimensions
- Purpose & Capability
- okThe name/description promise a pre-publish checklist tool and the provided artifacts implement that: SKILL.md documents checklist commands and examples; scripts/check.sh outputs checklist templates for code/api/deploy/version/launch/regression; scripts/script.sh implements a small local CLI and simple local data storage. No unrelated binaries, credentials, or cloud access are requested.
- Instruction Scope
- okRuntime instructions (SKILL.md) direct the agent to produce and use checklists and example CLI commands. The included scripts do not instruct the agent to read arbitrary system files, secrets, or network endpoints; check.sh formats checklist output, and script.sh reads/writes a local data directory only. There is no vague 'gather whatever context you need' instruction that would broaden scope.
- Install Mechanism
- okNo install specification is present (instruction-only with bundled scripts). No remote downloads or archive extraction are performed. The only code shipped is the two small shell scripts and docs, so nothing arbitrary is fetched at install time.
- Credentials
- okThe skill declares no required environment variables or credentials. The scripts use common environment variables (TEST_PUBLISH_CHECK_DIR optional override, XDG_DATA_HOME, HOME) to choose a local data directory — this is proportional and expected for a local CLI. No SECRET/TOKEN/PASSWORD vars are requested or accessed.
- Persistence & Privilege
- okalways:false and normal autonomous invocation. The skill writes small files to a per-user data directory (default: ${XDG_DATA_HOME:-$HOME/.local/share}/test-publish-check) — a reasonable, scoped filesystem presence. It does not modify other skills or system-wide agent settings.