The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection from processed technical documents.
Ingestion points: The skill reads content from vision.md, requirements.md, and spec.md in the project directory, as well as a guide file located in the user's home directory (~/.local/share/dotfiles/ai/guides/feature-development-process.md). These documents are external to the skill and potentially attacker-controlled.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard or escape instructions found within the ingested markdown files.
Capability inventory: The agent is instructed to generate, write, and update several files, including plan.md, plan-backend.md, plan-frontend.md, and discussion-summary.md based on the contents of the input documents.
Sanitization: The skill lacks mechanisms to sanitize, validate, or filter the content of ingested files before they influence the agent's logic and the resulting output files.

feature-plan