create-pipeline-v1

Pass

Audited by Gen Agent Trust Hub on Mar 18, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: Indirect prompt injection surface detected. The skill accepts user instructions to generate pipeline YAML which is then uploaded via the harness_create tool. * Ingestion points: User requirements provided as natural language to the agent (documented in SKILL.md). * Boundary markers: Absent. No instructions are provided to the agent to treat user-provided scripts or strings as untrusted or to delimit them within the YAML output. * Capability inventory: The harness_create MCP tool in SKILL.md allows writing configurations directly to the Harness platform. * Sanitization: Absent. The skill does not instruct the agent to sanitize, validate, or escape user input before interpolation into the generated YAML.
  • [EXTERNAL_DOWNLOADS]: Documentation reference to an external, non-vendor repository. * Evidence: references/v1-spec-schema.md references https://github.com/thisrohangupta/spec for the pipeline schema specification, which is a personal repository outside the official vendor (harness) organization.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 18, 2026, 12:11 PM
Security Audit — agent-trust-hub — create-pipeline-v1