The Agent Skills Directory

[SAFE] (SAFE): The skill consists entirely of Markdown-based instructions and prompts. No executable code, shell scripts, or obfuscated content were found.
[Indirect Prompt Injection] (LOW): The skill identifies surfaces for untrusted data ingestion, specifically in topic-review.md (processing user-provided drafts/links) and the referenced topic-research (processing external search results).
Ingestion points: topic-review.md receives external text or file content; topic-research (described in SKILL_OLD.md) ingests search engine results.
Boundary markers: Absent; the prompts do not explicitly use delimiters to separate untrusted data from instructions.
Capability inventory: The skill is limited to generating Markdown text and providing advice. It lacks the capability to write to the filesystem, execute system commands, or perform network operations beyond standard search tool usage.
Sanitization: None detected.
Assessment: Due to the lack of high-privilege side effects (no file-write or code-exec), the severity is low.
[Command Execution] (SAFE): The skill references a use_skill tool. Analysis indicates this is an internal state-management mechanism used to transition between planning modules (e.g., command="topic-scene-guide") rather than an interface for arbitrary shell command execution.

Topic_Planning