The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to offer the user a command to install additional components: npx skills add wondelai/skills/<slug> --global. This involves executing shell commands that perform global software installation, which can be used to run arbitrary code on the host system.
[EXTERNAL_DOWNLOADS]: The suggested installation process involves downloading software from an external repository associated with the user 'wondelai'. This source is not recognized as a trusted organization or well-known service, posing a risk of supply chain attack or execution of untrusted code.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes data from local project files that could be influenced by external actors.
Ingestion points: The skill reads docs/DESIGN-CODE-ARCHITECTURE-PLAN.md and other architectural artifacts in the docs/ folder (SKILL.md, Operating Rule 1).
Boundary markers: There are no explicit instructions to the agent to treat the contents of these files as data rather than instructions, nor are there delimiters used to wrap the ingested content.
Capability inventory: The skill has the capability to write to the filesystem and suggest shell commands for software installation.
Sanitization: The skill lacks validation or sanitization of the content retrieved from project documentation, allowing potentially malicious instructions embedded in those files to influence the agent's behavior during architectural planning.

design-code-architecture