The Agent Skills Directory

[NO_CODE] (SAFE): The skill consists exclusively of Markdown documentation and reference files (SKILL.md, references/checklist.md, references/frameworks.md). There are no scripts, binaries, or package manifests included.
[COMMAND_EXECUTION] (SAFE): No shell commands, subprocess calls, or administrative operations were detected in the instructions.
[DATA_EXFILTRATION] (SAFE): No network operations (curl, wget, fetch) or external domain references are present. There is no mechanism to send data outside the local environment.
[CREDENTIALS_UNSAFE] (SAFE): No API keys, tokens, or hardcoded secrets were found in the text.
[PROMPT_INJECTION] (SAFE): The instructions are strictly focused on providing a framework for document review and do not contain any patterns intended to bypass AI safety filters or override system constraints.
[DATA_SURFACE] (LOW): As a document review skill, it naturally processes user-provided data (indirect prompt injection surface). However, since the skill possesses no capabilities (network, file-write, or shell access), any embedded instructions in the reviewed documents cannot lead to security compromise.

doc-review