The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides shell scripts (scripts/spawn-implementer.sh and scripts/check-env.sh) that execute various system commands, including git, python3, and specialized tools like wt (worktrunk) and orca-ide. These scripts are used to manage worktrees, parse JSON output, and interface with the orchestration engine.
[INDIRECT_PROMPT_INJECTION]: The script scripts/spawn-implementer.sh reads task specifications from an external file provided via the --spec-file argument and passes the raw content directly into the orca-ide orchestration command.
Ingestion points: The --spec-file parameter in scripts/spawn-implementer.sh allows the skill to load arbitrary text data from a file on the file system.
Boundary markers: No explicit delimiters or instruction isolation techniques (such as "ignore embedded instructions" headers) are used when passing the spec file content to the worker sessions.
Capability inventory: The skill has the capability to initialize new worker sessions, manage git worktrees, and execute orchestration commands that can influence subsequent agent behaviors.
Sanitization: There is no evidence of sanitization, validation, or escaping of the ingested specification file content before it is processed.

dispatch-implement