The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from local project files which are considered untrusted input. Maliciously crafted content in these files could theoretically influence agent behavior during the modeling process.
Ingestion points: product/index.yaml, product-context.yaml, and story descriptions.
Boundary markers: No specific delimiters or instructions are used to isolate ingested data from the agent's logic.
Capability inventory: The skill has the ability to execute shell commands (npx, git) and write to the filesystem.
Sanitization: The instructions do not specify any validation or sanitization for the ingested text.
[EXTERNAL_DOWNLOADS]: The skill instructs the agent to use npx for YAML validation, which downloads tools from the npm registry at runtime.
Evidence: SKILL.md Step 4 and references/yaml-guardrails.md contain commands using npx js-yaml and npx tsx.
Source: Downloads originate from the npm registry, a well-known service.
[COMMAND_EXECUTION]: The skill executes various shell commands for file management, validation, and version control.
Evidence: SKILL.md Step 4 includes a script to validate YAML files and git commands to commit/push changes.

aep-model