docslime-adr

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses a local CLI tool named docslime to generate new ADR files and manages an index log via standard shell commands. It also uses grep for verifying tasks.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes project documentation, including requirements and existing ADR files, to help the agent contextually fill in new records.
  • Ingestion points: Reads from docs/engineering/adrs/, docs/3-ENGINEERING/ADRs/, and ../../REQUIREMENTS.md.
  • Boundary markers: Not explicitly defined for raw file reads.
  • Capability inventory: Local file creation/writing via the docslime CLI and standard shell redirection.
  • Sanitization: Relies on user interview steps and manual verification to validate the content of the ADRs.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 07:33 PM
Security Audit — agent-trust-hub — docslime-adr