delivery-record-verify
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from markdown files which represents an injection surface.
- Ingestion points: The skill reads and parses markdown files located at
docs/delivery-records/*.mdand potentially other user-provided paths. - Boundary markers: There are no explicit instructions or delimiters defined to isolate the content of the delivery records from the agent's instructional context.
- Capability inventory: The agent has the capability to execute local scripts (
python3,bash) and read files from the repository. - Sanitization: No sanitization or filtering of the file contents is described; the skill relies on a script to parse YAML frontmatter and evaluate the body content against threshold rules.
- [COMMAND_EXECUTION]: The skill invokes local executable scripts as its primary workflow.
- Evidence: It executes
python3 scripts/delivery_record_verify.pyandscripts/delivery-record-verify.shwith arguments derived from user input ([path ...]). While these are bundled scripts, they interact with the file system and process inputs that could be manipulated if the underlying scripts do not properly sanitize arguments.
Audit Metadata