The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands using node -e, python3 -c, and jq to parse transcript data. It also uses execSync within local Node.js scripts (fetch-slack-users.mjs, slack-post.mjs) to perform API requests via curl.
[EXTERNAL_DOWNLOADS]: The skill makes network requests to the Slack API (https://slack.com/api/) using curl to fetch user lists and post messages. These interactions are with a well-known service and are part of the skill's intended functionality.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8). It ingests untrusted meeting transcripts from Fireflies and passes this content to subagents without sufficient boundary markers or sanitization. Malicious instructions embedded in a transcript could potentially influence the agent's behavior when generating Linear ticket proposals or Slack DMs. Ingestion points: Fireflies transcripts fetched via MCP. Boundary markers: Missing in Phase 3 subagent prompts for transcript chunk processing. Capability inventory: Linear ticket creation and modification via MCP; Slack DM posting via local scripts. Sanitization: No evidence of input sanitization or validation of transcript content before processing.

extract-my-action-items