autumn-catalog

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the atmn CLI tool to perform operations such as project initialization (atmn init), configuration synchronization (atmn pull), and pushing validated updates to the server (atmn --headless push --yes). These commands are standard for the tool's integration and include instructions for manual review of diffs.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to transform natural language pricing descriptions into TypeScript configuration files. While this creates a surface for prompt injection, the risk is mitigated by the skill's instructions to perform human-in-the-loop reviews before any configuration is pushed.
  • Ingestion points: User-provided pricing details and requirements within the chat interface (SKILL.md).
  • Boundary markers: The skill employs a structured progress checklist and 'STRICT RULES' to guide the agent's processing of user input.
  • Capability inventory: The agent can write to autumn.config.ts and execute system commands through the atmn CLI tool.
  • Sanitization: The workflow relies on manual user verification of the generated catalog and configuration diffs prior to deployment.
  • [DATA_EXFILTRATION]: The skill interacts with the vendor's official API at api.useautumn.com through the atmn CLI and documented SDK examples. This data transfer is a required part of the skill's functionality to manage and sync pricing models.
  • [EXTERNAL_DOWNLOADS]: The documentation references the use of official vendor packages such as atmn, autumn-js, and autumn_sdk from standard package registries. These are recognized as legitimate resources belonging to the skill's authoring organization.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 10:49 PM
Security Audit — agent-trust-hub — autumn-catalog