learn-codebase
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill uses directive and forceful language such as "critical and non negotiable" and "no matter how many there are" to ensure the agent performs the exhaustive reading task, which acts as a behavioral override.
- [INDIRECT_PROMPT_INJECTION]: The skill mandates that the agent ingest the full content of every source file in the project. This significantly increases the risk of the agent processing malicious instructions embedded in codebase files (e.g., in comments or documentation) without adequate safety boundaries.
- Ingestion points: Every source file within the project codebase via the
Readtool, as instructed inSKILL.md. - Boundary markers: None specified in the instructions to separate untrusted codebase content from the agent's core instructions.
- Capability inventory: The skill utilizes the
Readtool to access file contents. - Sanitization: No sanitization, filtering, or validation of the codebase content is performed before the agent processes it.
Audit Metadata