linkedin-post-formatter

Pass

Audited by Gen Agent Trust Hub on Aug 4, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill includes defensive instructions that mandate treating all user-provided source material as data rather than instructions, effectively preventing malicious content from overriding the agent's core logic.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data provided by the user. 1. Ingestion points: user-provided ideas, facts, excerpts, and transcripts (SKILL.md). 2. Boundary markers: The 'Safety and factual boundaries' section explicitly instructs the agent to treat quoted text as data and not instructions. 3. Capability inventory: Analysis of the skill instructions and referenced files shows no tool access or capabilities for command execution, network operations, or file system modifications. 4. Sanitization: The skill enforces a strict policy of non-execution for any instructions discovered within processed data.
  • [OBFUSCATION]: A Unicode character map is provided in references/unicode-charmap.md for styling purposes. While these characters are technically homoglyphs, the skill includes specific safeguards by requiring a plain-text version be provided first and warning the user about accessibility risks, which prevents deceptive use.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 4, 2026, 02:06 AM
Security Audit — agent-trust-hub — linkedin-post-formatter