hook-generator
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill uses emphatic directives such as 'CRITICAL: Auto-start on load' and negative constraints ('Do not summarise', 'Do not explain', 'No preamble') to override the agent's default conversational guidelines and safety wrappers regarding verbosity.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data (topics) which are interpolated into the prompt flow. Although the skill has no dangerous capabilities, the lack of sanitization creates a surface for indirect instructions.
- Ingestion points: User-supplied 'topic' collected in Step 1.
- Boundary markers: None present to isolate user-supplied text from the agent's core instructions.
- Capability inventory: None detected; the skill is limited to generating text hooks.
- Sanitization: No escaping or filtering is applied to the input topic.
- Multi-step chain: Step 4 explicitly instructs the agent to suggest calling another skill ('post-formatter'), creating a potential injection chain if the output hook contains malicious instructions for the next skill.
Audit Metadata