The Agent Skills Directory

[COMMAND_EXECUTION]: The skill involves the execution of several shell-based development tools and local scripts:
Development tools: git status, git diff, ruff, pytest, and python manage.py are utilized for PR validation and conflict resolution.
Extension management: The skill instructs the use of a local Python script scripts/install-external-skills.py to add external skill chains.
Setup commands: The installation instructions include npx and file system operations (mkdir, cp) to deploy the skill.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves processing external, potentially untrusted data which presents an injection surface:
Ingestion points: PR diffs, descriptions, comments, linked issues, and tool outputs are processed during triage and review workflows.
Boundary markers: The protocol lacks explicit instructions to use delimiters or ignore instructions embedded in the PR content, making the agent susceptible to following commands hidden in the data.
Capability inventory: The skill possesses the ability to write to the file system (resolving merge conflicts), stage changes, and execute installation scripts.
Sanitization: No explicit sanitization, validation, or filtering of the ingested PR data is performed before processing.
[EXTERNAL_DOWNLOADS]: The documentation points to the author's GitHub repository (github.com/alvarovillalbaa/plugins) for the skill installer and refers to external registry files (external-skills.yaml, external-sources.yaml).

prs