Skip to main content
skill CX Codex

feedback — Claude Code skill: reports and installation

This Claude Code skill builds a report about using PLSI and can email it to the PLSI team. It activates on feedback requests, after setup, and during the weekly feedback routine.

MIT ★ 93 updated 2026-11-15

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add Insilico-org/plsi --skill feedback -a claude-code
Open on GitHub

What feedback does

feedback collects frustrations, improvement ideas, and stage-time estimates from the chat and PLSI thread ledgers. The plsi feedback report command checks the draft and creates files for the user and the email; the skill then presents the report and asks before sending. Nothing is emailed until the user chooses Send.

The workflow needs PLSI and its plsi feedback timeline and plsi feedback report commands. Data Controller checks the report before delivery. Email comes from the user's Grok account and goes to PLSI's feedback address; if that address is unset, the report is saved without an email. The report excludes personal details and medical-record contents.

What it helps with

  • compile setup frustrations from a PLSI chat
  • estimate time spent on work stages
  • turn feedback into concrete PLSI suggestions
  • send a reviewed report to the PLSI team

Example prompts for the agent

  • Summarize what frustrated me while setting up PLSI.
  • Estimate how long I spent on each PLSI stage and prepare a report.
  • Send the PLSI team my feedback and improvement ideas.

How to install feedback

1Skills CLI (Vercel)

npx skills add Insilico-org/plsi --skill feedback -a claude-code

2Manually: put the files in place

~/.claude/skills/feedback/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/feedback/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about feedback

How do I install feedback in Claude Code?

Run npx skills add Insilico-org/plsi --skill feedback -a claude-code. The skill file is installed at ~/.claude/skills/feedback/SKILL.md.

What does feedback do?

It turns chat and PLSI thread history into a usage report, shows the draft, and emails it only after the user confirms Send.

What does feedback need to work?

It needs PLSI with the plsi feedback timeline and plsi feedback report commands. Data Controller checks the report, and sending requires the user's Grok email account plus a configured PLSI feedback address.

Source text

SKILL.md · MIT
---
name: feedback
description: Use as PLSI when the user wants to tell the PLSI team how PLSI works for them - "compile a usage report", "what frustrated me in setup", "how long did setup take", "send feedback", "suggest improvements" - and at its offers (end of setup, then the weekly "PLSI feedback" routine) to write a usage report with the frustrations, the time per stage and suggested improvements, and send it from the user's Grok email to the Longevity Mothership's feedback address.
---

# PLSI usage report

The report is about PLSI, never about the user's health: it goes to the PLSI team by email, so it needs no Mothership
join and no token. Data Controller clears it before the user sees the send card (`shared/team` rule 6).

## When to use

- The user asks: "compile a usage report", "list what frustrated me", "how long did each step take", "send
  feedback to the developers", "what should the devs improve?".
- **Offers**, only when "Feedback on PLSI" is in the user's Mothership choice (setup 6.7, `orchestrator/mothership`
  step 1; kept in your permanent memory as "Feedback offers: on"), each as ONE short card "Help improve PLSI? I can
  send the PLSI team a short report of <what>.": See the report / Not now:
  - end of setup, once the record card is answered (`plsi.md` step 14): "how setup went". Right after it, create
    the ENABLED weekly routine **"PLSI feedback"** (an existing one is updated to this);
  - each "PLSI feedback" run: "your week with PLSI", the report built from this week's chats, thread ledgers
    (`orchestrator/conduct`, `failed` rows included) and `feedback mark/timeline`. The run stays quiet when nothing
    was marked that week (no stage timed, no frustration or `failed` row), or after two "Not now" in a row; then
    offer again only when the user brings it up or a month after the second "Not now".
  Keep the offers made and the "Not now" count in your permanent memory; a See the report resets the count. The user
  can always ask.

## Steps

1. **Compile.** From this chat and the thread ledgers (`orchestrator/conduct`), write `<work>/feedback-draft.json`:
   - `frustrations`: each moment the user was slowed, confused or annoyed (a question asked twice, a wait, an error,
     a card that did not fit, an answer they had to correct, a step they asked to skip), as {`stage`, `what`,
     `severity`}: `high` blocked or lost work, `medium` cost minutes or a retry, `low` a small annoyance;
   - `suggestions`: one concrete change for each frustration that has one, and the user's own ideas, as {`stage`,
     `what`};
   - `stages`, optional: estimated minutes for use stages (`records`, `plans`, `research`) as {`stage`, `minutes`,
     `measured`: false}; setup stages are timed by code.
   `stage` is one of `install`, `team`, `choices`, `interview`, `specialists`, `connectors`, `instance`, `routines`,
   `mothership`, `welcome`, `records`, `plans`, `research`, `other`. Each `what` is one plain sentence of at most 300
   characters, about what PLSI did: no names, health values, record content, dates, addresses or links. At most 20
   of each. Then:

   ```bash
   plsi feedback timeline
   plsi feedback report --file <work>/feedback-draft.json
   ```

   `report` adds the timed setup stages, checks the message and writes `usage-report_<YYYYMMDD-HHMMSS>.md` (the
   user's copy) and `.json` (the message). `GATE_FAILED`: rewrite the named entries and run it again.

2. **Clear.** A `request` to Data Controller (`controller/egress`) with the `.json` in `files` and "feedback report"
   in `summary`. `rejected`: rewrite the entries it names and run step 1 again.

3. **Show and ask.** ONE message with the `.md` from `report` attached: the total setup time and the longest
   stage, the top 3 frustrations and the top 3 suggestions, in plain words. Then ONE card "Send this report to the
   PLSI team? It goes from your Grok email.": Send / Edit / Just keep it. The user asked only for the report (no
   "send"): the same card. `email` is null (no feedback address set): no card; say the report is saved.

4. **Send.** Send: email `email.to`, `email.subject` and `email.body` exactly as drafted, from the user's Grok email;
   one line that it went; the routine **"PLSI Mothership replies"** is created or updated by the primary hub
   (another hub asks it hub to hub; `shared/hub`, Mothership replies): the Mothership's receipt or refusal arrives in the user's Grok email and wakes it. Edit: ask
   what to change in one chat line, apply it to the draft, then steps 1-3 again. Just keep it: one line with the file
   name.

## Rules

- Nothing is sent without the Send answer on that card; an offer is never a send.
- The report is written from what happened in PLSI, not from the user's records; a frustration that only makes sense
  with health data is written without it ("a lab report was read wrong", never the value).
- Never ask the user to fill in the report: they see the result and can edit it.

© Insilico-org · MIT · Insilico-org/plsi

Similar skills

All skills for Agent setup: memory, context, orchestration →

From the same set85

The whole set →
hallmarks This Claude Code skill maps a gene, pathway, or candidate to aging hallmarks and supporting evidence. It activates for requests to assess those links and possible biomarkers. CX Codex +1 ★ 93repo stars company Builds sourced research dossiers on pharma and biotech companies. It runs for a company name or ticker, or before pharma/outlook when the dossier is older than 7 days. CX Codex +1 ★ 93repo stars landscape The landscape Claude Code skill finds drugs in development against a gene. It runs on demand during target discovery or evaluation when structured sources leave gaps. CX Codex +1 ★ 93repo stars l-llm The l-llm skill for Claude Code and Codex routes requests to Longevity-LLM through PLSI. It focuses on candidate genes and targets linked to aging. CX Codex +1 ★ 93repo stars add-map The add-map Claude Code skill adds a PLSI map for a measured quantity that installed maps do not cover. It activates when a bot reports a map gap or a user asks to track a new quantity. CX Codex +1 ★ 93repo stars patient This Claude Code skill routes a physician’s work on one patient to the appropriate workers. It handles incoming files, visit preparation, notes, evidence reviews and plan options. CX Codex +1 ★ 93repo stars ward The ward Claude Code skill manages individual patient bots in a PLSI team. It creates one when a physician starts work on a roster patient, or removes it after the patient leaves the roster. CX Codex +1 ★ 93repo stars review This Claude Code skill checks PLSI plan predictions against follow-up results and revises decisions. It runs when a review is due or new results arrive. CX Codex +1 ★ 93repo stars

Project sponsors

Companies that keep this analytics open