dd-triage-flaky-test

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: Executes local shell commands git and the pup CLI (DataDog's CI Visibility tool) to retrieve repository context and test failure telemetry.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting untrusted data from test logs which are then processed by the agent.
  • Ingestion points: Extracts @error.message and @error.stack from DataDog search results in STEP 2 of the workflow.
  • Boundary markers: The skill does not use specific delimiters or instructions to encapsulate the ingested test failure data as untrusted content.
  • Capability inventory: Includes subprocess execution (git, pup) and file system writes (/tmp/flaky-update.json).
  • Sanitization: No sanitization or escaping is performed on the ingested error messages or stack traces before they are presented to the agent for analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 01:51 PM
Security Audit — agent-trust-hub — dd-triage-flaky-test