link-validation

Pass

Audited by Gen Agent Trust Hub on Nov 7, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to use bunx or brew to fetch the lychee link checker from public registries. Lychee is a well-known open-source tool. * Evidence: Instructions to use 'bunx lychee' and 'brew install lychee'.
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute link validation tasks and repository management commands. * Evidence: Use of 'git rev-parse --show-toplevel', 'moon run', and 'bash tasks/link-check.sh'.
  • [PROMPT_INJECTION]: The skill contains instructions for 'Self-Evolving' behavior, directing the agent to modify the SKILL.md file itself based on runtime outcomes. * Evidence: The 'Self-Evolving Skill' header and 'Post-Execution Reflection' section.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes markdown files within the user's workspace, which acts as an ingestion point for potentially untrusted data. * Ingestion points: Markdown files (./**/*.md) processed during validation. * Boundary markers: None present. * Capability inventory: Access to Bash, Read, and Glob tools. * Sanitization: No sanitization is performed on the content of the files being checked for links.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 7, 2026, 02:12 AM