The Agent Skills Directory

[OBFUSCATION]: The file scripts/dev-workflow.cjs is a large, minified, and bundled JavaScript file (nearly 4,000 lines of minified code). This practice effectively obfuscates the script's internal logic, decision-making processes, and potential side effects, making it extremely difficult to verify the safety of the script through static analysis.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external, potentially untrusted sources like checkpoint.md, 00-master-plan.md, and other PRD files.
Ingestion points: Content is read from various files within the .dev/ directory to synthesize a resumption summary.
Boundary markers: The instructions do not specify the use of delimiters or clear boundary markers to isolate the ingested file content from the agent's primary instructions.
Capability inventory: The agent has access to the Bash tool for arbitrary command execution and the Read tool. The custom CLI also has the ability to read and write files and manage symbolic links.
Sanitization: There is no mention of sanitizing or filtering the content read from external Markdown files before it is presented to the LLM for synthesis.
[COMMAND_EXECUTION]: The skill performs shell operations via scripts/discover.sh and scripts/git-state.sh, and utilizes child_process.execFile within the dev-workflow.cjs script to interact with the system's git binary and file system. While these are necessary for the skill's functionality, they represent a significant attack surface if input validation were to be bypassed.

dev-resume