The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted repository-local data, such as .coderabbit.yaml and CI configuration files, to audit and propose changes. This creates a surface for indirect prompt injection where instructions hidden within these files could attempt to manipulate the agent's behavior.
Ingestion points: Reads .coderabbit.yaml, repository instructions, and CI files (referenced in SKILL.md).
Boundary markers: The instructions do not define clear delimiters or specific isolation protocols for the ingested data.
Capability inventory: The agent can read local files and propose writes/edits to repository configuration files.
Sanitization: The skill relies on schema validation for YAML but does not explicitly sanitize the text content for embedded instructions.
[EXTERNAL_DOWNLOADS]: The skill references and fetches configuration schemas and documentation from CodeRabbit's official domains to assist in the audit and validation process.
Evidence: References https://coderabbit.ai/integrations/schema.v2.json, https://docs.coderabbit.ai/reference/configuration, and https://docs.coderabbit.ai/configuration/path-instructions in SKILL.md.

config-review-bot