eo-recall
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from the local project environment, including source code, design documents, and research notes, to generate explanations. This creates a surface for indirect prompt injection if those files contain adversarial content.\n
- Ingestion points: Reads from
state/,decisions/,brainstorm/,research/, and source code files.\n - Boundary markers: The instructions require the agent to cite specific sources (file:line or document path) and explicitly state if information is missing from documentation.\n
- Capability inventory: File system read access, execution of the
codegraphutility for symbol exploration, and file system write access to thetmp/eo/explain/directory for HTML reports.\n - Sanitization: The skill does not describe specific sanitization steps for the ingested code or document text.\n- [COMMAND_EXECUTION]: The workflow involves running
codegraph exploreto analyze code structures and call chains. This is an expected behavior for a code-recall tool but involves executing a sub-process to interact with the project repository.
Audit Metadata