harness-engineering-lifecycle
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides installation instructions that use a package manager to fetch resources from the author's registry (
arenukvern/skill_steward). As this points to the author's own infrastructure, it is consistent with the skill's deployment model. - [COMMAND_EXECUTION]: The instructions guide the agent to perform local builds, install development versions of harnesses, and execute validation suites against consumer repositories. This involves running shell commands to establish baselines and verify integration across project boundaries.
- [PROMPT_INJECTION]: The skill establishes an attack surface for indirect prompt injection by instructing the agent to ingest and process data from external repository sources.
- Ingestion points: The agent is directed to read
steward.yamlcontracts and validation outputs from sibling repositories (SKILL.md Part 2). - Boundary markers: The instructions do not define specific delimiters or instructions to ignore potentially malicious content embedded in the ingested configuration files.
- Capability inventory: The agent is authorized to perform shell-based builds, installations, and test executions based on the state of the ingested repository data.
- Sanitization: There is no mention of sanitizing or validating the contents of the external configuration files before processing.
Audit Metadata