The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the gh (GitHub CLI) tool to check authentication status (gh auth status) and create issues (gh issue create). These commands are limited to the vendor's specific repository (Arkiv-Network/reported-issues) and require explicit user confirmation before execution.
[DATA_EXFILTRATION]: While the skill submits data to a public GitHub repository, this is its primary intended function. The skill mitigates accidental data exposure by requiring a full preview and confirmation from the user, and by instructing the agent to redact common secrets (private keys, mnemonics) found in user-provided logs.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input (bug descriptions, logs) and incorporates it into a drafted issue body. This surface is mitigated by the structured markdown template and the mandatory human-in-the-loop confirmation step before any submission occurs.

arkiv-feedback