The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it retrieves and processes untrusted user-generated content from the Farcaster network.
Ingestion points: The script scripts/neynar.sh fetches external content in cmd_user, cmd_feed, cmd_search, and cmd_cast functions.
Boundary markers: No explicit delimiters or instructions are used to warn the agent to ignore embedded commands in the fetched content.
Capability inventory: The skill possesses write capabilities including posting casts, liking content, and following users.
Sanitization: While the script uses jq to extract and format specific fields, the underlying text content is not sanitized for malicious instructions targeting the agent.
[COMMAND_EXECUTION]: The script executes curl and jq to interact with the Neynar API. It correctly uses shell variable quoting and jq's argument-passing features (such as --arg) to prevent shell injection vulnerabilities when handling user-provided strings.
[DATA_EXFILTRATION]: The skill performs network operations to api.neynar.com, which is the official endpoint for the Neynar service. It reads API credentials from a local configuration file (~/.clawdbot/skills/neynar/config.json) and sends them to this endpoint for authentication, which is consistent with the primary purpose of the skill. The script uses heredocs and standard input for POST data to avoid exposing sensitive data in process listings.

neynar