The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill utilizes instruction chaining by directing the agent to load and execute a protocol defined in a sibling skill path (../alignfirst/). This is a modular design pattern for coordinating multiple skills.
Ingestion points: The instructions in SKILL.md reference local files at ../alignfirst/SKILL.md and references/description-protocol.md.
Boundary markers: None present.
Capability inventory: No system commands, network access, or file-system write operations are present in this file.
Sanitization: Not applicable as the skill does not ingest untrusted remote data.
[PROMPT_INJECTION]: The skill contains a directive to "Do not use your plan mode." This is a platform-specific instruction used to force the agent into immediate execution of the referenced protocol. In this context, it serves as a workflow control rather than an attempt to bypass safety guidelines.

aldescription