The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze untrusted external data, which creates a surface for indirect prompt injection attacks.
Ingestion points: As described in SKILL.md, the agent reads "any artefact the agent can read" and briefing documents provided by the delphi-ground skill.
Boundary markers: The skill requires a structured output format and grounding labels (observed, inferred, constructed), providing some semantic distinction, but it lacks formal technical delimiters (e.g., XML tags or special fences) to isolate external input from the agent's instructions.
Capability inventory: According to the frontmatter in SKILL.md, the skill has access to Read, Write, Grep, Glob, and Bash tools.
Sanitization: The instructions do not define any sanitization, filtering, or escaping mechanisms for the content ingested from external documents.
[NO_CODE]: The skill consists entirely of Markdown instructions and documentation without including any executable scripts or binary files.

delphi-imagine