The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to retrieve and analyze data from Honcho workspaces, peer cards, and session transcripts which may contain untrusted user content.
Ingestion points: Data is pulled into the agent context via honcho session view, honcho workspace search, and honcho conclusion list commands.
Boundary markers: The instructions do not specify the use of delimiters or boundary markers for the data retrieved from the Honcho platform.
Capability inventory: The skill has access to Bash (scoped to honcho and jq), Read, and Grep tools.
Sanitization: There are no explicit instructions to sanitize, escape, or validate the content retrieved from the external workspace before the agent processes it.
[CREDENTIALS_UNSAFE]: The documentation identifies ~/.honcho/config.json as the storage location for the apiKey. While this is standard configuration for the tool, documenting the specific path to sensitive credentials increases the risk of accidental exposure if the agent is prompted to read or display configuration files.
[EXTERNAL_DOWNLOADS]: The honcho start command initiates the download of Docker images to set up a local stack. These resources are associated with the vendor's primary functionality.
[COMMAND_EXECUTION]: The skill utilizes a suite of honcho CLI commands to perform administrative and diagnostic tasks, including environment configuration and workspace management.

honcho-cli