Skip to content
Agentmart
Submit a tool

upsum

Close a work session by appending a durable record, rebuilding a fixed-size recency-weighted summary, carrying open work forward, and running repository checks.

Skill by lilCnid

IndexedCC-BY-4.0
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/lilcnid-upsum-cb1e0f

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Close a work session by appending a durable record, rebuilding a fixed-size recency-weighted summary, carrying open work forward, and running repository checks. Invoke explicitly.

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found
Runs as
Python package

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/upsum

What's inside

Skill name
upsum
Description
Close a work session by appending a durable record, rebuilding a fixed-size recency-weighted summary, carrying open work forward, and running repository checks. Invoke explicitly.
License
CC-BY-4.0
Instructions
95 lines in SKILL.md · view on GitHub

Before installing

  • Ships executable scripts

Files (4)

  • LICENSE.md
  • SKILL.md
  • openai.yaml · agents/
  • checks.py · scripts/

SKILL.md (first lines)

# Upsum

The record grows and never forgets. The summary stays bounded and forgets on
purpose. Run this ceremony only when explicitly invoked.

## Locate the project memory

Use existing repository conventions when present. Otherwise use:

| Path | Rule |
|---|---|
| `.upsum/RECORD.md` | append-only session record |
| `.upsum/SUMMARY.md` | rewritten whole, 250-word default budget |
| `TODO.md` | rewritten whole, human-visible open work |

## 1. Append one evidence-based record entry

Read the diff and repository state, then append:

```md
## {ISO_Date} · {Short_Session_Label}

**Did** — {Repository_Changes_Not_Conversation_Intent}
**Learned** — {New_Evidence_Or_Understanding}
**Decided** — {Decision_And_Owner}
**Left** — {Deliberately_Unfinished_Work_And_Reason}
```

Never edit earlier entries. If no work changed and no durable decision was made,
say so and stop instead of adding noise.

## 2. Rebuild the bounded summary

Derive `.upsum/SUMMARY.md` from the entire record. Keep the configured budget,
or 250 words when none exists. Allocate resolution by recency:

```text
oldest  {Collapsed_To_A_Label}
         {Outcome_Only}
         {Outcome_Plus_One_Distinguishing_Detail}
newest  {Nearly_Full_Detail_With_Redundancy_Trimmed}
```

Entries age by losing resolution, not by disappearing. Do not grow the budget
because the record grew.

## 3. Carry forward open work

Rewrite `TODO.md` from the record and current session. Each open item names the
blocking decision or condition and who can unblock it:

```md
- [ ] {Observable_Completion} — blocked on {Owner_And_Condition}
```

Remove completed items. `Left` records intentional incompletion; it is not an
apology field.

## 4. Run the shipped checks

Resolve this skill's directory from its source locator in the current skill
catalog, then run its script against the repository working directory:

```bash
python {upsum-skill-directory}/scripts/checks.py {repository-directory}
```

Do not assume the skill lives under a global config directory; plugins may run
from a repository scope or versioned cache. The script checks inside references,
repository state, skill health, and license/credit travel. `UNMEASURED` means the
check could not establish a result, not that it passed.

Exit codes describe measurement health:

- `0`: every check ran;
- `1`: at least one check could not be measured;
- `2`: the checker itself failed.

Act on findings or record why a finding is accepted. Do not merely narrate the
output.

## 5. Refresh publishing claims

Only when the session will commit, push, or release:

- test the documented install path from a clean location;
- verify that the README inventory matches the repository;
- verify that every stated limit still holds.

## Completion

Report which memory files changed, the summary budget, open blockers, check
results, and any unmeasured evidence. Preserve the append-only record exactly.

From lilCnid/dovetail-codex, CC-BY-4.0.

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Sep 3, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
CC-BY-4.0
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from lilCnid

Cite this page

upsum skill. Agent Mart. https://agentmart.dev/skills/lilcnid-upsum-cb1e0f (last updated 2026-10-24)