chronicle-protocol
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill consists entirely of descriptive instructions and configuration metadata. It does not ship with any executable scripts, binaries, or automated installation steps.
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to read and write project state from external files, creating a potential surface for indirect prompt injection if checkpoint files are tampered with.
- Ingestion points: The agent reads files following the pattern
chronicle_<project-name>_<YYYY-MM-DD>.mdfrom the project directory. - Boundary markers: The protocol explicitly requires the assistant to summarize the contents of a retrieved checkpoint and wait for user confirmation before acting on the data.
- Capability inventory: The instructions direct the agent to utilize available file-write, file-read, and shell tools to persist and retrieve data.
- Sanitization: The skill relies on the assistant's summarization and the user's manual review as a verification gate for external content.
Audit Metadata