article-prompts-to-skills

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill uses untrusted external data (articles and tutorials) as input for generating instructions and code, creating a surface for indirect prompt injection. Malicious source content could attempt to manipulate the extraction process or the resulting skill files.
  • Ingestion points: SKILL.md (Step 1) reads source articles and tutorials to identify reusable prompts.
  • Boundary markers: Steps 2 and 3 in SKILL.md define strict boundaries and extraction rules to separate portable mechanics from source-specific branding.
  • Capability inventory: The skill is instructed to write files, manage version control via git, and execute a local Python script.
  • Sanitization: Step 7 in SKILL.md includes a specific instruction to scan changed files for secrets, tokens, and private paths.
  • [COMMAND_EXECUTION]: The skill utilizes local shell commands to perform its tasks. It instructs the agent to use git status, git diff, and git commit for repository management, and to run a local script quick_validate.py to verify the new skill packages.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 01:37 PM
Security Audit — agent-trust-hub — article-prompts-to-skills