The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill instructs the agent to read sensitive API keys from a local file located at ~/.gooseworks/credentials.json.
[COMMAND_EXECUTION]: Uses python3 -c to programmatically extract data from local files and set environment variables, providing a shell execution surface.
[REMOTE_CODE_EXECUTION]: Instructs the user to run npx gooseworks login, which downloads and executes code from the NPM registry at runtime.
[DATA_EXFILTRATION]: Transmits user-supplied lead data (emails and names) to the external domain api.gooseworks.ai for enrichment, which is the stated purpose of the skill but involves sending potentially sensitive PII to a third party.
[EXTERNAL_DOWNLOADS]: Performs outbound network requests to api.gooseworks.ai and api.github.com to fetch enrichment data and social proof signals.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because user-provided inputs (email and name) are interpolated into shell commands (curl) and JSON payloads without sanitization or boundary markers.
Ingestion points: email and name parameters in SKILL.md.
Boundary markers: Absent.
Capability inventory: Shell command execution via curl and python3 in SKILL.md.
Sanitization: No validation or escaping is applied to user data before it is included in command strings.

gtm-enrichment-smart