skills/paldom/skillskit/publish-repo/Gen Agent Trust Hub

publish-repo

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill automates repository publishing tasks using standard CLI tools including git, gh (GitHub CLI), npx, and curl. These operations include flipping repository visibility, managing API-based security settings, and creating GitHub releases.
  • [EXTERNAL_DOWNLOADS]: The skill uses npx skills@latest, which downloads and executes the ecosystem's CLI tool from the NPM registry. This is standard behavior for the toolset described.
  • [DATA_EXFILTRATION]: During the 'Verify like a consumer' step, the skill explicitly unsets privacy-related environment variables (DISABLE_TELEMETRY, DO_NOT_TRACK) to ensure the installation is reported to the skills.sh service. This 'seeding' is necessary for the repository to appear in the public catalogue.
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes untrusted data from the local skills.sh.json file to determine repository groupings and descriptions. Ingestion points: skills.sh.json (Step 1 and 3). Boundary markers: Absent in command execution. Capability inventory: gh repo edit, gh api, gh skill publish, npx skills. Sanitization: No specific sanitization or validation of the JSON content is performed before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 09:48 PM
Security Audit — agent-trust-hub — publish-repo