rust-lifecycle

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands to manage branches and validate the project environment. Evidence: SKILL.md and references/workflow-steps.md use git for branch management and gh for pull request operations.
  • [REMOTE_CODE_EXECUTION]: The skill invokes specialized sub-agents to perform implementation and review tasks. Evidence: references/workflow-steps.md demonstrates the use of the Task() tool to launch agents within the rust-agents namespace.
  • [DATA_EXFILTRATION]: The skill transmits project code and metadata to remote repositories as part of its core functionality. Evidence: references/workflow-steps.md uses git push and gh pr create to update GitHub repositories with phase completion artifacts.
  • [PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection through its agent-to-agent handoff protocol. (1) Ingestion points: Data from sub-agents enters the workflow via YAML files in the .local/handoff/ directory. (2) Boundary markers: Absent; handoff file paths are interpolated into sub-agent prompts without explicit delimiters or instructions to ignore untrusted content. (3) Capability inventory: The orchestrator possesses the ability to execute git/gh commands and launch further agents with the Task() tool. (4) Sanitization: Absent; while the orchestrator filters specific fields for logic control, the complete handoff file is accessible to the next agent in the chain without sanitization.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 09:55 AM
Security Audit — agent-trust-hub — rust-lifecycle