The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides examples for linting agent configurations and running scripted calls within CI environments. These activities involve the agent ingesting and interpreting external data, which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Agent configuration files, test scripts, and call transcripts processed in CI workflows described in SKILL.md and references/ci-configs.md.
Boundary markers: The skill does not define specific delimiters or "ignore embedded instructions" warnings for the processed data.
Capability inventory: The skill is permitted to use Bash (restricted to npm and curl), Write, Edit, Read, and Grep tools.
Sanitization: While the documentation suggests masking credentials and redacting transcripts, there is no technical implementation for sanitizing or escaping the content of the configuration files to prevent prompt-based attacks.

retellai-ci-integration