oma-debug
Pass
Audited by Gen Agent Trust Hub on May 16, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it is designed to analyze untrusted external data such as error messages, logs, and reproduction steps.\n
- Ingestion points: External data enters the agent's context through logs, stack traces, and user-provided reproduction steps as noted in
SKILL.md.\n - Boundary markers: The instructions lack explicit boundary markers or delimiters to separate untrusted data from the system prompt (Absent).\n
- Capability inventory: The agent has extensive capabilities including executing shell commands for testing and linting, as well as the ability to read and write files within the codebase as noted in
SKILL.mdandresources/execution-protocol.md.\n - Sanitization: No evidence of sanitization or validation of external input before processing (Absent).\n- [COMMAND_EXECUTION]: The skill uses various shell commands (e.g.,
rg,npm test,pytest) to perform its primary function. These tools, while necessary for debugging, could be misused if the agent is compromised by malicious instructions embedded in the data it processes.
Audit Metadata