The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from pull request diffs and descriptions which could contain malicious instructions designed to influence the agent's analysis.
Ingestion points: Pull request bodies and diff contents fetched via gh pr diff and gh pr view (SKILL.md).
Boundary markers: The skill does not explicitly use delimiters or specific safety instructions when passing diff content to the analysis logic.
Capability inventory: The skill can execute shell commands (gh, glab, grep, jq), perform network requests (curl), and run test suites (npm test, pytest) (SKILL.md).
Sanitization: Ticket identifiers are extracted using regular expressions before being used in API calls, providing some mitigation for the ticket verification step.
[COMMAND_EXECUTION]: The skill executes local test runners (npm test, pytest) to calculate coverage deltas. While this is standard for a code review workflow, it involves executing code present within the pull request being reviewed.
[EXTERNAL_DOWNLOADS]: The skill interacts with external APIs to verify project management tickets.
Evidence: Performs network requests to atlassian.net (Jira) and api.linear.app (Linear) using curl.
Note: These are well-known services and the skill uses environment variables (JIRA_API_TOKEN, LINEAR_API_KEY) for authentication rather than hardcoded secrets.

pr-review-expert