sop-library

Pass

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local command-line tool using user-provided input. * Evidence: The 'CHECK-COMPLIANCE' process step calls 'execution-tracker.py --brand {slug} --action list'. * Risk: If the '{slug}' variable (Brand slug) is not strictly validated, it could serve as a vector for command injection attacks.
  • [DATA_EXFILTRATION]: The skill interacts with the local filesystem to manage SOP and brand data. * Evidence: Read and write operations occur in '/.claude-marketing/brands/' and '/.claude-marketing/sops/'. * Context: This behavior is aligned with the skill's function to store and retrieve agency standard operating procedures.
  • [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes external JSON data. * Ingestion points: The skill reads SOP content from '/.claude-marketing/sops/' and brand profiles from '/.claude-marketing/brands/'. * Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present when processing this data. * Capability inventory: The skill has the ability to write files to the local directory and execute the 'execution-tracker.py' script. * Sanitization: There is no documented sanitization or validation of the contents of the SOP or brand profile files before they are processed.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 27, 2026, 01:03 PM