The Agent Skills Directory

[SAFE]: The skill consists entirely of natural language instructions and YAML-formatted data structures for logic modeling. It does not include any scripts, command execution, or network operations.
[NO_CODE]: No Python, Node.js, or shell scripts are present in the skill definition.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data describing system processes and failure modes. While this creates a potential surface for indirect prompt injection if the source data is untrusted, the skill does not possess any high-risk capabilities (such as file writing or network access) that could be abused via such an injection.
Ingestion points: Input contract in SKILL.md defining functions_or_process, failure_modes, and observed_effects.
Boundary markers: Absent.
Capability inventory: None (no tool calls, network, or file system access).
Sanitization: Absent.

build-failure-chain