skills/tumf/cflx-skills/cflx-proposal/Gen Agent Trust Hub

cflx-proposal

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes a local Python script at scripts/cflx.py to handle specialized operations like listing, showing, and validating Conflux proposals.
  • [COMMAND_EXECUTION]: Shell commands rg (ripgrep) and ls are used for context gathering and repository exploration, which is essential for drafting relevant proposals.
  • [PROMPT_INJECTION]: The skill processes repository content that could contain indirect prompt injections.
  • Ingestion points: Data is ingested from existing specifications and repository files through rg and ls commands during the research phase.
  • Boundary markers: The skill does not use specific delimiters or instructions to ignore embedded instructions in the context it gathers.
  • Capability inventory: The skill can write files to the openspec/changes/ directory and execute the cflx.py helper script.
  • Sanitization: The cflx.py script performs structural validation using regular expressions but does not filter out natural language instructions from the processed text.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 02:49 AM