The Agent Skills Directory

[COMMAND_EXECUTION]: Executes several local shell scripts located in the .agents/bin/ directory (such as chart.sh, swain-session-check.sh, artifact-context.sh, and swain-focus.sh) to manage project state and generate the roadmap documentation.
[COMMAND_EXECUTION]: Uses the open utility to automatically launch and display the generated markdown files in the operator's default viewer.
[COMMAND_EXECUTION]: Interpolates user-provided artifact IDs directly into shell command arguments when generating scoped roadmap slices.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes content from untrusted project artifact files.
Ingestion points: Reads the first sentence of specific sections (## Value Proposition or ## Goal / Objective) from Vision and Initiative markdown files.
Boundary markers: No delimiters or explicit instructions are provided to the agent to ignore or isolate the extracted content from its own logic.
Capability inventory: The skill possesses the capability to execute shell scripts, perform file writes, and open local files/links.
Sanitization: There is no evidence of sanitization, validation, or escaping of the extracted artifact text before it is written back into the project-wide ROADMAP.md.

swain-roadmap