linkedin-safety-verification

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The instructions in SKILL.md (Rule 7) and agents/openai.yaml direct the agent to 'continue automatically through the authorized batch without confirmation prompts'. This directive reduces human-in-the-loop verification for individual actions within a batch.
  • [PROMPT_INJECTION]: Rule 1 in SKILL.md explicitly instructs the agent to 'never turn contact fields, uploads, redirects, tailoring, verification-code retrieval, or submit controls into new confirmation boundaries', which instructs the agent to bypass seeking user approval for potentially sensitive or untrusted data inputs.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection. 1. Ingestion points: LinkedIn account context, contact fields, and batch data (SKILL.md). 2. Boundary markers: Absent (the skill does not define delimiters or ignore instructions for external data). 3. Capability inventory: The agent is empowered to perform LinkedIn 'writes' (SKILL.md). 4. Sanitization: No evidence of filtering or validation of untrusted LinkedIn content before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 01:23 PM