The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data sources such as logs and stack traces, which creates a surface for indirect prompt injection attacks.
Ingestion points: The skill parses logs and error patterns as described in the focus areas of SKILL.md;
Boundary markers: The instructions do not provide guidance on using delimiters or instructions to ignore commands that may be embedded in the logs;
Capability inventory: The skill is used for searching logs and codebases, which requires read access to various files within the environment;
Sanitization: No sanitization or filtering of the extracted log content is mentioned in the instructions.

error-detective