The Agent Skills Directory

[DYNAMIC_EXECUTION]: Several validation and utility scripts, including validate-skill.sh, sync-plugin-manifest.sh, validate-evals.sh, and migrate-licensing.sh, execute hardcoded Python logic via the python3 <<'PYEOF' pattern. This is used for advanced data processing tasks such as JSON manipulation and complex regex validation that exceed standard shell capabilities.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to perform repository management operations. As evidenced in the scripts/ directory, these actions include using git for version control, gh for GitHub integration, and glab for GitLab workflows. These commands are necessary for the skill's intended purpose of automating repository maintenance and releases.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from external repository files, including SKILL.md markdown and evals.json structured data, through its validation scripts (scripts/validate-skill.sh and scripts/validate-evals.sh). These scripts perform structural audits and regex-based checks on the ingested data to ensure compliance with repository standards.
[EXTERNAL_DOWNLOADS]: The documentation references external resources from trusted domains. For example, references/validation-checklist.md provides an example command for downloading a scanner from the NousResearch/hermes-agent GitHub repository. Additionally, CI validation prompts (SR-34 and SR-38) suggest fetching configuration or guidelines from the skill's own official repository on GitHub.

skill-repo