penguinradar/skill
Developer tools이 페이지는 AI로 자동 번역되었습니다. 기술적인 세부 사항은 원문을 확인하세요. 원문 보기
Ask in plain language; the assistant reads your retained follower and following snapshots and detected changes through PenguinRadar's CLI or MCP tools, and explains what the data does and does not...
개요
Ask in plain language; the assistant reads your retained follower and following snapshots and detected changes through PenguinRadar's CLI or MCP tools, and explains what the data does and does not show. - "Which Instagram accounts am I tracking?" - "Show alice's last available followers." - "What changes did PenguinRadar detect for alice in the last seven days?" - "Who stopped following alice since October 1?" - "Query bob's public followers online, at most 2 credits." This installs the Skill into the assistants found on your machine. With the PenguinRadar CLI you can install the same Skill without GitHub: penguinradar skill install (--target claude for ~/.claude/skills, --project for one project). The Skill contains no credentials and installs no software. Give the assistant access one of two ways: 1. add https://penguinradar.com/mcp as a connector and sign in on PenguinRadar's own authorization page. 2.
README
원문 README의 일부입니다. 원문 보기.
PenguinRadar Agent Skill
An Agent Skill that lets Claude Code, Codex, Cursor and other compatible AI assistants answer questions about the Instagram accounts you track with PenguinRadar. Ask in plain language; the assistant reads your retained follower and following snapshots and detected changes through PenguinRadar’s CLI or MCP tools, and explains what the data does and does not show.
Example prompts
- “Which Instagram accounts am I tracking?”
- “Show alice’s last available followers.”
- “What changes did PenguinRadar detect for alice in the last seven days?”
- “Who stopped following alice since October 1?”
- “Query bob’s public followers online, at most 2 credits.”
Install
npx skills add penguinradar/skill
This installs the Skill into the assistants found on your machine. With the PenguinRadar CLI you can install the same Skill without GitHub: penguinradar skill install (--target claude for ~/.claude/skills, --project for one project).
Connect your account
The Skill contains no credentials and installs no software. Give the assistant access one of two ways:
- Remote MCP (OAuth): add
https://penguinradar.com/mcpas a connector and sign in on PenguinRadar’s own authorization page. - CLI:
npm install --global penguinradar, thenpenguinradar auth login(macOS Keychain) or setPENGUINRADAR_API_KEY.
Never paste an API Key into a chat.
What the assistant can do
| You ask about | MCP tool | CLI command |
|---|---|---|
| Tracked accounts | list_tracking |
penguinradar tracking list --json |
| Account profile and totals | get_account |
penguinradar account USERNAME --json |
| Existing followers | list_followers |
penguinradar followers USERNAME --json |
| Existing following | list_following |
penguinradar following USERNAME --json |
| Detected changes | get_account_changes |
penguinradar changes USERNAME --json |
| Online lookup of a public account | query_public_account |
penguinradar followers USERNAME --live --max-credits N --json |
Built-in safeguards
- Existing data first. Reads use data PenguinRadar already holds and never start a scan.
- Explicit budgets. An online lookup runs only after you name the target, the data type and a maximum number of credits. It returns one page and never starts continuous tracking.
- Honest data. The assistant reports the data’s observation time, coverage and warnings. “Detected at” is not the exact time someone followed or unfollowed, a partial list is not the whole Instagram list, and no detected change is not proof of no activity.
- Data is not instructions. Names and bios returned by the API are treated as data only.
FAQ
추천 도구
다른 키워드를 입력하거나 필터를 제거해 보세요.
설치
npx skillfish add penguinradar/skill