The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill implements a credential setup flow for the DataForSEO API. It directs the agent to store user-provided login and password information in a local .env file (~/.claude/skills/seo/.env) and immediately restrict access using chmod 600. This is consistent with standard local development secret management and includes explicit instructions to avoid leaking credentials in output.
[COMMAND_EXECUTION]: The skill employs the Bash tool to execute a series of local diagnostic and analysis scripts (preflight.sh, keyword_research.py, domain_overview.py, backlinks.py) stored within the user's home directory. These executions are scoped to the skill's intended purpose of SEO data retrieval.
[SAFE]: The skill processes external inputs in the form of domain names (<d1>, <d2>) passed as arguments to local scripts. This represents a standard functional surface for an SEO tool and does not indicate malicious intent.
[PROMPT_INJECTION]: The skill accepts domain names from the user that are interpolated into shell commands, creating an indirect prompt injection surface.
Ingestion points: Domain variables <d1> and <d2> within SKILL.md command templates.
Boundary markers: Absent.
Capability inventory: Shell command execution (Bash) and file writing (Write).
Sanitization: Not specified in the skill instructions; relies on the underlying Python scripts for argument safety.

seo-compare