api-testing-skill
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONPERSISTENCE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and process data from external API endpoints, which could contain malicious instructions designed to influence the agent's behavior during test evaluation.\n
- Ingestion points: API response bodies are parsed using
pm.response.json()inreference/postman-patterns.mdandres.bodyinreference/bruno-patterns.md.\n - Boundary markers: There are no specific delimiters or instruction-ignore markers defined for the data being validated.\n
- Capability inventory: The skill utilizes shell-based CLI execution for Newman and Bruno (
reference/ci-integration.md) and enables JavaScript execution within the testing tool environments for assertions and request chaining.\n - Sanitization: The patterns do not include sanitization steps for external response data before it is interpolated into environment variables or logged.\n- [PERSISTENCE]: The skill directs the agent to write a telemetry file to the user's home directory after execution, persisting usage metrics across sessions.\n
- Evidence:
SKILL.mdcontains an instruction to write the final outcome to~/.claude/skill-analytics/last-outcome-api-testing-skill.json.\n - Context: This is a non-executable logging mechanism used for analytics and tracking skill performance metrics.
Audit Metadata