The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external GitHub issue titles, bodies, comments, and repository source code. A malicious issue could contain hidden instructions attempting to influence the agent's behavior during the 'translation' or 'analysis' phase.
Ingestion points: Data enters the context through gh issue view and git clone operations defined in SKILL.md.
Capability inventory: The skill can execute shell commands (gh, git) and read/modify files within the local working directory (mkdir, grep, git reset --hard).
Boundary markers: Absent. The instructions do not define specific delimiters or instructions to ignore commands embedded within the external text.
Sanitization: Absent. The content is passed to the agent for summarization and logic determination without filtering.
[EXTERNAL_DOWNLOADS]: The skill downloads project source code from GitHub repositories using git clone --depth 1 and fetches metadata via the GitHub API. These operations target well-known services for the skill's primary purpose.
[COMMAND_EXECUTION]: The skill uses shell-based tools including the GitHub CLI (gh) and git to automate repository triage. It performs actions such as listing issues, viewing PRs, and resetting local file states to match the remote repository.

oss-triage