refresh-reports
Skill0 starsChanged 12 days agoMITView on GitHub
Rerun every report on fresh data and republish each page to its existing link. Run by the Monday scheduled task in Claude Desktop, or by hand with /refresh-reports.
In this repository6 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- aerobic-progressSkill
- load-checkSkill
- race-predictorSkill
- trail-predictorSkill
- weekly-reviewSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 3 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 14
Run npm run export first, every time.
- Very specific: line 21
Run each of these skills in full, in this order, through to publishing its page: weekly-review, load-check, race-predictor, …
- Distinctive: line 24
Always write new notes from today's numbers, even when notes already exist.
- Very specific: line 26
Read metrics.json with jq or by line range: it's too large to read whole.
- Distinctive: line 27
Keep the write-ups short: the page is the result of this run.
- Distinctive: line 36
If one page fails, leave its published version alone, note why, and carry on with the next.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Refresh reports
This run is usually unattended: nobody can answer a question, so never wait for one. Where a skill says to ask the runner, put the question in the page's notes and carry on.
1. Check the data
Run npm run export first, every time. If it fails, publish nothing and end with the last line
below: the published pages stay as the last good ones. Then read until and version from
data/metrics.json (jq '{until, version}' data/metrics.json): if until isn't today or
version isn't 1, stop the same way.
2. The pages with a skill
Run each of these skills in full, in this order, through to publishing its page: weekly-review,
load-check, race-predictor, aerobic-progress, trail-predictor.
- Always write new notes from today's numbers, even when notes already exist. Rebuilding from old notes publishes old words over new numbers.
- Read
metrics.jsonwithjqor by line range: it's too large to read whole. - Keep the write-ups short: the page is the result of this run.
3. The pages without a skill
For race-form and training-breaks: run npm run report -- <page>, read its link from what it
prints, read that artifact, then publish data/reports/<page>.html to that link.
4. Failures
If one page fails, leave its published version alone, note why, and carry on with the next. Never publish a page whose build failed, and never compute a number to fill a gap.
5. Finish
List each page with its link or its failure. The last line of your reply must be exactly
REFRESH OK if all 7 pages were published, otherwise REFRESH FAILED: <the pages that failed>.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

