The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from project documentation and passes it to subagents for code generation, creating a surface where instructions within those files could influence the agents behavior.\n
Ingestion points: Reads task lists and design docs from paths such as docs/work/YYYY-MM-DD-<slug>/tasks.md and plan.md.\n
Boundary markers: The skill does not instruct the agent to wrap ingested content in delimiters or include 'ignore instructions' warnings when dispatching subagents.\n
Capability inventory: The skill has access to Bash, Write, and Edit tools, allowing for filesystem modification and command execution.\n
Sanitization: There is no evidence of sanitization or schema validation for the ingested markdown content.\n- [COMMAND_EXECUTION]: The skill executes local shell scripts and project-specific CLI tools as part of its development workflow.\n
Evidence: Execution of local conflict detection scripts via bash scripts/conflict-detect.sh and interaction with the forge CLI for synchronization and progress logging. These operations are within the expected scope for a tool designed to manage development tasks.

dev