token-optimizer
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill's primary function involves processing untrusted external text, which establishes an indirect prompt injection surface. * Ingestion points: Data is ingested via scripts/compress.py and scripts/compress_prompt.py through standard input or file paths. * Boundary markers: Output is structured using markdown headers as defined in SKILL.md. * Capability inventory: The skill is limited to local regex-based string transformations with no network access or unsafe execution capabilities. * Sanitization: Security-critical data is protected by regex patterns and a defined 'Never Compress' list in references/never_compress.md to prevent corruption of credentials and identifiers.
Audit Metadata