The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes a local shell script scripts/init-project.sh to initialize the directory structure and create markdown template files for new projects. The script is provided within the skill package and only performs benign file system operations.
[PROMPT_INJECTION]: The skill is designed to synthesize insights from untrusted user-provided content, such as interview transcripts and survey exports, which are stored in the research/raw/ directory. This creates an indirect prompt injection attack surface.
Ingestion points: The agent reads raw text files from research/raw/ for synthesis and clustering.
Boundary markers: The instructions do not specify the use of clear delimiters or markers to isolate the ingested data from the agent's instructions.
Capability inventory: The skill can execute shell scripts, write files to the project directory, and perform network operations for desk research.
Sanitization: The skill does not explicitly describe sanitization or filtering of the raw data content, though it includes a 'Judge' role to verify that insights are grounded in evidence.
[EXTERNAL_DOWNLOADS]: The documentation references the installation of the skill and its optional companion, market-researcher, using the command npx skills add tronghieu/agent-skills. These resources originate from the vendor's repository and are part of the standard installation and workflow process.

design-thinking