The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several CLI tools to manage the repository and deployment lifecycle.
Evidence: Executes git ls-files, git commit, and git push for version control management. It uses bun run build and bun run deploy for application building and Cloudflare Workers deployment, as well as curl for verifying production status.
[EXTERNAL_DOWNLOADS]: The skill references external package execution via npm.
Evidence: The skill instructs the use of npx skills@latest to manage skill additions and installations from the public registry.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing external metadata from the skill collection.
Ingestion points: Automatically enumerates and reads frontmatter from all skills/*/SKILL.md files in the repository to generate index content and blog data.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present when processing this data.
Capability inventory: The skill has the ability to write to files (app/data.ts), execute build scripts, and push to remote repositories.
Sanitization: There is no logic provided to sanitize or escape potentially malicious instructions embedded in the summary or description fields of the skills being indexed.
[DATA_EXFILTRATION]: The skill accesses sensitive local paths and performs network operations consistent with its primary purpose as a development and release tool.
Evidence: Accesses files within ~/Documents/GitHub/cankun-blog and ~/Documents/GitHub/cankun-skills to sync metadata and project state. Changes are pushed to GitHub repositories, which is expected behavior for an automated release manager.

skill-release