digest · insilico-org — Claude Code skill: what it does, how to install
This Claude Code skill gathers new research for a PLSI digest or an on-demand scan. It connects verified findings to active plans, gaps, and goals.
MIT ★ 93 updated 2026-11-15
npx skills add Insilico-org/plsi --skill digest -a claude-codeWhat digest does
digest searches for recent material on topics drawn from active plans, open gaps, goals, and the request itself. It filters by publication date, skips items from earlier digests, and checks each paper with plsi evidence search and plsi evidence verify.
The skill writes a digest file and sends PLSI two sections: a user-facing reading list and findings tied to plans or goals. The list can contain up to 5 items; news and industry reports are kept distinct from clinical evidence. digest does not edit plans or contact the user directly. It needs PLSI and the plsi command, and it cannot create the recurring digest routine itself.
What it helps with
- scan recent research on a chosen topic
- verify identifiers for scientific papers
- connect new findings to an active plan
- find evidence related to a gap or goal
Example prompts for the agent
- Run a research scan on sleep studies published since the last digest.
- Search PubMed for findings related to an active plan.
- Find new evidence for the open goals and send the results to PLSI.
How to install digest
1Skills CLI (Vercel)
npx skills add Insilico-org/plsi --skill digest -a claude-code2Manually: put the files in place
~/.claude/skills/digest/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/digest/
1Skills CLI (Vercel)
npx skills add Insilico-org/plsi --skill digest -a codex2Manually: put the files in place
.agents/skills/digest/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add Insilico-org/plsi --skill digest -a cursor2Manually: put the files in place
.cursor/skills/digest/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install Insilico-org/plsi skills/explorer/digest2Manually: put the files in place
.github/skills/digest/SKILL.md1Skills CLI (Vercel)
npx skills add Insilico-org/plsi --skill digest -a gemini-cli2Platform CLI
gemini skills install https://github.com/Insilico-org/plsi --path skills/explorer/digestManually: put the files in place
.devin/skills/digest/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about digest
How do I install digest in Claude Code?
Run npx skills add Insilico-org/plsi --skill digest -a claude-code. The skill file is installed at ~/.claude/skills/digest/SKILL.md.
What does digest do?
It searches for and verifies recent research, then sends PLSI a digest and findings connected to active plans, gaps, or goals.
What does digest need to run?
It needs PLSI and the plsi command. Searches draw on digest and journal sources, as well as PubMed.
Source text
SKILL.md · MIT
---
name: digest
description: Use as PLSI Explorer when PLSI's digest routine or the user asks for a research scan - the longevity digest for the user and new insights for the plan (findings that bear on active plans, open gaps or goals).
---
# PLSI digest and insights
## When to use
- PLSI's "PLSI longevity digest" routine sends a `request` (thread `digest-<date>`, the start date in `summary`).
- PLSI asks for an ad-hoc scan on a topic.
- Never create a routine yourself; the digest routine runs on PLSI.
## Steps
1. Read the sources and what the user is working on (no records, only published states and plans):
```bash
plsi refs list
plsi status --instance <instance>
plsi maps select --instance <instance>
plsi state show --instance <instance> --map personal_context
```
2. Pick 3-7 topics, in this order: what active plans target, open gaps and out-of-range coordinates, goals, the
selected maps; a topic named in the request comes first. Search the `digest` and `journal` sources and PubMed for
items on them published since the date in the request, and drop items an earlier digest already listed (the
`digest_*.md` files in `<PLSI_HOME>/work/digest-*/`). Queries use topics only, never personal data:
```bash
plsi evidence search --source pubmed --query "<topic>"
```
3. Verify every paper before listing it; drop what does not verify:
```bash
plsi evidence verify --kind <pmid|doi> --id <id>
```
4. Write the digest items with their identifiers to `<PLSI_HOME>/work/<thread>/digest_<YYYYMMDD>.md` (today's date,
`shared/team` rule 3) and send one `result` to PLSI on the request's thread, with that file in `files` and two
parts in `summary`:
- **Digest** for the user: at most 5 items, each with title, link, one line on why it matters to this user, and its
kind (research, event, industry report). Industry reports and events are news, not clinical evidence.
- **Insights for the plan**: verified findings that bear on an active plan (a prediction's basis, a new risk, a
stronger or contradicting trial) or on a gap or goal, each with the plan id or coordinate and the identifier.
Nothing new: a `result` saying "Nothing new" (PLSI stays quiet with the user).
## Rules
- Never message the user or another bot directly; PLSI shows the digest and forwards the insights to PLSI Reviewer
(`reviewer/review`) for active plans and to PLSI Nostradamus (`nostradamus/plan`) for a new option.
- Never change plans or record anything.
© Insilico-org · MIT · Insilico-org/plsi
Similar skills
All skills for Science and research →digest by other authors comes first.