The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to run git and gh (GitHub CLI) commands for repository management. It executes gh auth status, git tag, git show, gh issue view, and gh issue comment. These commands are standard for the skill's stated purpose and are executed within the local repository context.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted data from external sources. 1. Ingestion points: The skill reads issue titles, bodies, and comments via gh issue view in Step 5 and parses CHANGELOG.md in Step 3. 2. Boundary markers: No explicit boundary markers are used to isolate ingested content from instructions during comment generation. 3. Capability inventory: The skill has the capability to write to GitHub issues using gh issue comment in Step 8. 4. Sanitization: While no automated sanitization is mentioned, the skill implements a mandatory 'HARD STOP' preview and user approval step before any comments are posted. This human-in-the-loop verification effectively mitigates the risk of malicious instructions in issues being automatically executed or propagated.

jaan-issue-solve