The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes a local binary located at .slipbox/bin/slipbox to perform configuration lookups and note validation. This execution is restricted to the specific relative path within the managed workspace, intended for system-specific tasks like retrieving paths and casing conventions.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from Literature notes and Resource files to perform synthesis, which creates an attack surface for indirect prompt injection if those sources contain malicious instructions.
Ingestion points: Reads content from Literature notes wikilinked as 'Key Concepts' and resolves original Resource identities.
Boundary markers: Relies on an upstream '/grounding' process and maintains an audit trail in synthesis-map.json to track contributing sources and evidence.
Capability inventory: Performs file read/write operations and executes workspace-specific validation tools.
Sanitization: Employs /write-checks to evaluate draft style and humanize signals, and enforces a mandatory post-write validation step using the slipbox note validate command.

make-reference-note