The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard development tools, specifically git and the GitHub CLI (gh), to resolve commit SHAs and fetch CI run information. Evidence includes calls to git rev-parse HEAD in SKILL.md and scripts/watch-ci.mts, and gh repo view in scripts/wait-for-pr-event.mts.
[EXTERNAL_DOWNLOADS]: The skill facilitates the download of its own source files from its official GitHub repository for installation and configuration purposes. The README.md file provides installation commands targeting the github.com/Vivswan/skills repository.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data by fetching CI logs and Pull Request comments/reviews, creating an ingestion surface for potentially malicious instructions embedded in external content.
Ingestion points: scripts/watch-ci.mts retrieves CI job logs via gh run view --log-failed, and scripts/wait-for-pr-event.mts monitors PR comments and reviews using GraphQL queries.
Boundary markers: The skill's instructions in SKILL.md direct the agent to report fetched content but do not specify the use of delimiters or instruction-bypass warnings to mitigate risks from data content.
Capability inventory: The skill operates within an environment capable of executing shell commands and accessing the file system via the Bun runtime.
Sanitization: External content from logs and comments is reported to the agent as excerpts without specific sanitization or filtering of instruction-like patterns.

watch-ci-after-push