The Agent Skills Directory

[PROMPT_INJECTION]: The skill implements a persistent learning mechanism where the agent is instructed to append newly discovered tips or workarounds to references/learnings.md. This creates a surface for indirect prompt injection if malicious instructions are provided by a user and stored in the reference file for future sessions.
Ingestion points: User descriptions of problems and platform interactions (Step 1 and Step 4 in SKILL.md).
Boundary markers: Absent. The skill does not define specific delimiters for separating user-provided data from system instructions when writing to the learning file.
Capability inventory: File-write operations to references/learnings.md and tool routing capabilities to other /sales-* skills.
Sanitization: Absent. There is no instruction to validate or sanitize content before it is persisted to the local file system.
[COMMAND_EXECUTION]: The documentation includes instructions for users to install related tools using the command npx skills add sales-skills/sales --skill sales-do. This command references a package within the vendor's own namespace.
[DATA_EXFILTRATION]: The skill provides detailed instructions on how to authenticate with the LeadIQ GraphQL API (https://api.leadiq.com/graphql). While it does not contain hardcoded credentials, it describes the process of encoding API keys found in user account settings into HTTP Basic Authentication headers.

sales-leadiq