The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and refine information from external, potentially untrusted sources including work artifacts, pull requests, and issue descriptions. When combined with the skill's capabilities to prepare pull requests and perform code merges or deployments, this creates a surface for indirect prompt injection where malicious instructions embedded in external data could influence the agent's actions.
Ingestion points: The skill reads from work artifacts, local Markdown, codebases, pull requests, and deployed results as specified in SKILL.md.
Boundary markers: Although the 'Original Request' section is separated by a horizontal rule in references/work-artifact.md, there are no specific instructions for the agent to treat this section as untrusted or to ignore instructions contained within it.
Capability inventory: The skill is authorized to refine code, prepare pull requests, and execute merges, releases, or deployments upon authorization as per SKILL.md.
Sanitization: No explicit sanitization or filtering of external content is described in the provided skill files.

equilio-integrate