odin
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by passing unsanitized user input directly to a sub-agent.
- Ingestion points: The
$ARGUMENTSvariable from the user command is utilized in Step 4 to construct the prompt for the Munin agent. - Boundary markers: Absent. The user input is appended directly to the instruction 'Search for information about: $ARGUMENTS' without the use of XML tags or other delimiters that might prevent the sub-agent from following instructions embedded within the user input.
- Capability inventory: The
bifrost-muninsub-agent is granted access to the<BIFROST_REPO>directory, allowing it to read files across several memory layers based on the search prompt. - Sanitization: No sanitization, validation, or escaping is performed on the
$ARGUMENTSstring before it is interpolated into the agent prompt.
Audit Metadata