The Agent Skills Directory

[Indirect Prompt Injection]: The skill is designed to ingest and process untrusted external data, such as git diffs, commit logs, and linked issue content, to generate pull request descriptions. This creates a surface where malicious content within a PR or issue could attempt to influence the agent's output or actions.
Ingestion points: The activation contract requires reading branch information, file diffs, commit logs, and issue descriptions (SKILL.md).
Capability inventory: The skill has the ability to write files (e.g., .git/PR_BODY.md) and execute shell commands for validation (SKILL.md).
Boundary markers: The instructions do not specify the use of delimiters or instructions to ignore embedded prompts in the analyzed data.
Sanitization: There are no explicit sanitization steps mentioned for the ingested data.
[Dynamic Execution]: The skill provides a shell script template for the agent to use when validating Mermaid diagrams. Executing shell commands based on templates is a common practice in developer tools but constitutes dynamic execution.
Evidence: SKILL.md contains a bash snippet using awk and npx to extract and validate diagrams.
[External Downloads]: The skill fetches the @mermaid-js/mermaid-cli package from the NPM registry to perform diagram validation.
Evidence: The validation step uses npx --yes -p @mermaid-js/mermaid-cli mmdc (SKILL.md).
Context: This utilizes a well-known service (NPM) and a widely recognized tool for its intended purpose.

pr-description-skill