The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx @ericrisco/rsc copy check <file> in SKILL.md to perform automated copy verification. This command downloads and executes a package from the author's own namespace.
[COMMAND_EXECUTION]: The skill includes and executes several local scripts for project maintenance and auditing:
scripts/verify.sh (Bash): Scans project files for banned words and structural issues using grep or ripgrep.
scripts/seo_audit.py (Python): Performs a technical SEO audit by fetching and analyzing a user-provided URL.
[DYNAMIC_EXECUTION]: The skill instructs the agent to run local shell and python scripts provided within the skill package to automate quality assurance and SEO auditing tasks.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection due to its brand grounding mechanism.
Ingestion points: Raw user-provided text and voice samples are saved verbatim to 02-DOCS/raw/brand/ as described in references/brand-grounding.md.
Boundary markers: The raw samples are stored with a simple provenance header but the content itself is unedited.
Capability inventory: The skill has the capability to write to the local filesystem (02-DOCS/), execute shell commands (npx), and perform network requests (via the SEO audit script).
Sanitization: There is no evidence of sanitization or escaping of the user-provided data before it is persisted or re-injected into the agent's context during the copy-writing phase.

marketing