guard-gauge-diagnostic

Pass

Audited by Gen Agent Trust Hub on May 3, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its primary function of processing untrusted external data.
  • Ingestion points: The skill explicitly accepts untrusted data from multiple sources including conversational dumps, messy notes, CRM tables, and uploaded PDF files, as defined in SKILL.md (Workflow Step 1) and references/input-rules.md (Rule 1).
  • Boundary markers: There are no instructions or system-level delimiters (e.g., XML tags, triple backticks, or specific 'end of data' markers) provided to separate the untrusted user content from the internal diagnostic instructions.
  • Capability inventory: The skill's capabilities are limited to text analysis, scoring, and generating diagnostic output. It does not have access to shell execution, network tools, or file system write operations.
  • Sanitization: No sanitization, escaping, or validation logic is specified to handle potentially malicious instructions embedded within the provided account data or PDF slides.
Audit Metadata
Risk Level
SAFE
Analyzed
May 3, 2026, 04:41 PM
Security Audit — agent-trust-hub — guard-gauge-diagnostic