Heard Product Feedback — Development skill for Claude Code
A claude skill to auto-collect feedback about your product from public sources.
How to install Heard Product Feedback
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open codingyogini/heard-product-feedback and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Heard Product Feedback does
A claude skill to auto-collect feedback about your product from public sources.
Alternatives in Development
- Survey — PUA 调研问卷 — 7 部分交互式问卷收集用户反馈 19.5k ★
- Iterate Implementation Plan — You are tasked with updating existing implementation plans based on user feedback 10k ★
- Metrics Dashboard — Define product metrics dashboard with sources and alert thresholds 7.8k ★
README
Heard — a product feedback Claude Skill
A Claude skill that collects **permitted public** product feedback, clusters it into themes, verifies every quote and count in code, and produces a self-contained HTML report with proposed **Build / Watch / Ignore** calls. The PM keeps the final decision.
By [Irene Pylypenko](https://irenepylypenko.com).
Install in Claude Code
/plugin marketplace add codingyogini/heard-product-feedback
/plugin install heard-product-feedback@irene-pylypenko
Use the skill directly (Claude.ai or other hosts)
Copy `plugins/heard-product-feedback/skills/heard-product-feedback/` into your skills folder, or zip that folder and upload it as a custom skill.
What it does
- Scope & source plan — confirms product, segment and approved public sources (RSS/Atom, App Store JSON, canonical JSON, manual import). No sign-in walls, no private CRM or support data.
- Collect & prepare —
scripts/heard.py collect/preparepulls records, strips author fields, dedupes and batches. - Analyze — the model assigns every record to themes or excludes it with a reason, and proposes a verdict per theme with counterevidence.
- Verify & report —
scripts/heard.py buildchecks quote substrings, IDs, coverage and intensity bounds, computes counts and ranking, and rendersreport.html. - Validate — drafts a bounded spec and falsifiable validation questions for the accepted top theme.
Licensed under MIT. See [PRIVACY.md](PRIVACY.md) for data handling.
Requires Python 3 and file access in the host. Tests: `cd plugins/heard-product-feedback/skills/heard-product-feedback/scripts && python -m unittest test_heard`.
Layout
.claude-plugin/marketplace.json # marketplace: irene-pylypenko
plugins/heard-product-feedback/
.claude-plugin/plugin.json
skills/heard-product-feedback/ # the skill (SKILL.md, scripts, references, assets)
Related Skills
Linear Curator Run
Run the curator — collect signals from configured signal sources and report
News Digest
Claude Code Skill: collect your own feeds, score novelty/significance/relevance, derive a reading priority fro
Enrich Person
Resolve the person at {{linkedin_url}} and verify employment at {{company_domain}} using public sources. Prese
Always Okay MCP
Creative-direction lens reconstructed from 125 public records of Min Hee-jin's work (interviews, talks, credit
Pc From Zero
Build a project's Product Context (product_context.md) from scratch by incrementally synthesizing sources the
Pc Update
Produce OLD→NEW find/replace edits that bring an existing product_context.md to its current decided state from
Related Agents
Council Graham
The product and startups seat on the Roast Room council. Reasons, decides, and talks the way Paul Graham does,
Nosy Neighbor
Researches a single rival product from public sources, writes/updates a file in the pm/rivals/_TEMPLATE.md sha
Council Blakely
The consumer product and brand from zero seat on the Roast Room council. Reasons, decides, and talks the way S