doc-brd-autopilot
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes content from external sources to generate documents.
- Ingestion points: Reference documents (
docs/00_REF/), user prompts (--prompt), and implementation plans (--iplan). - Boundary markers: There are no explicit instructions or delimiters defined to prevent the agent from following instructions embedded within the ingested data.
- Capability inventory: The agent can write files, execute local shell scripts, and run system commands.
- Sanitization: While it includes an audit cycle, there is no specific filtering for malicious natural language instructions.
- [COMMAND_EXECUTION]: The skill utilizes several standard shell utilities to process files and metadata. Evidence includes the use of
grep,ls,tr,cut, andsha256sumwithin described bash snippets for ID reservation and hash computation. - [REMOTE_CODE_EXECUTION]: The skill instructs the agent to execute a local validation script. Evidence: The execution of
bash ai_dev_ssd_flow/01_BRD/scripts/validate_brd_wrapper.shduring the audit phase.
Audit Metadata