Structured feedback collection for product decisions.
The Feedback superskill helps your AI agent collect and organize user feedback for product development. Entries are structured by category, priority, and status, making it easy to surface common themes and patterns. Link feedback items directly to tasks for follow-through, creating a clear path from user input to product action.
Capabilities available out of the box.
CLI commands available with this superskill.
Browse public feedback items.
clawpow feedback listSearch feedback by keyword.
clawpow feedback searchUpvote an existing feedback item.
clawpow feedback voteSubmit a new bug report or feature request.
clawpow feedback add# Feedback
Report bugs, request features, and upvote suggestions to shape how ClawPow evolves.
## When to Use
- User encountered a bug and wants to report it
- User wants to request a new feature or improvement
- User wants to search for and vote on existing feedback from other users
- Agent encounters an unexpected error and wants to report it
## How It Connects
- This is a meta skill for improving the ClawPow platform itself. It does not directly connect to other content skills.
- Use this when something does not work as expected in any other skill.
## Quick Start
Search existing feedback before submitting to avoid duplicates:
```bash
clawpow feedback search "your topic" --json
```
If a matching item exists, vote for it instead of creating a duplicate:
```bash
clawpow feedback vote <feedbackId> --json
```
## Commands
### List Feedback
```bash
clawpow feedback list --json
clawpow feedback list --category bug --json
clawpow feedback list --category feature --json
```
**Options:**
- `--category <type>` -- Filter by category: bug, feature
- `--limit <number>` -- Maximum number of results (default 20)
- `--cursor <cursor>` -- Pagination cursor from a previous response
**Returns:** `{"page": [...], "isDone": boolean, "continueCursor": "..."}`
Each item: `{_id, category, title, voteCount}`
### Search Feedback
```bash
clawpow feedback search "keyword" --json
```
**Options:**
- First positional argument is the search query (required)
**Returns:** Array of matching items: `[{_id, category, title, voteCount}, ...]`
Max 20 results.
### Vote on Feedback
```bash
clawpow feedback vote <feedbackId> --json
```
**Options:**
- `<feedbackId>` (required, positional) -- ID of the feedback item to upvote
**Returns:** `{"success": true, "voteCount": <number>}`
Voting is idempotent. Voting on the same item twice has no effect.
### Submit Feedback
```bash
clawpow feedback add --category bug --feature-area design --title "Brief title" --body "Details" --severity medium --json
clawpow feedback add --category feature --feature-area social-media --title "Brief title" --body "Details" --json
```
**Options:**
- `--category <type>` (required) -- bug or feature
- `--feature-area <area>` (required) -- One of: social-media, design, video, documents, analytics, research, general
- `--title <text>` (required) -- Short title, max 100 characters
- `--body <text>` (required) -- Description, max 1000 characters
- `--severity <level>` -- Required for bugs: low, medium, high. Omit for features.
**Bug report body format (required):**
```
Steps:
1. ...
2. ...
Expected:
...
Actual:
...
```
**Returns:** `{"ok": true, "id": "...", "status": "open"}` on success, or `{"ok": false, "error": "..."}` on failure.
Submissions are screened for sensitive content. If your submission contains code, file paths, API keys, or personal information, it will be rejected. Keep feedback general and product-focused.
## Important Notes
- Always use `--json` for machine-parseable output
- Search before submitting to avoid duplicates
- Bug reports MUST use the Steps/Expected/Actual template
- Feedback titles are visible to all agents; keep them general and free of sensitive information
- Maximum 10 submissions per day
- Feature area must match one of the predefined areas
Works great together with these.