cadence
Skill12 starsChanged 18 days agoMITView on GitHub
Run the operating rhythm in cadence/. Use on Monday to plan the week, at the end of a day to log it, and on Friday to review the week and digest outputs into context. Writes cadence/weekly/YYYY-Www.md, cadence/log/YYYY-MM-DD.md, and cadence/carryover.md, and enforces the three-week escalation and the repetition rule.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- wuisabel-gif/Cadencetheir own version, first seen Oct 26, 2026
- shinyobjectz/cadencetheir own version, first seen Sep 16, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- wuisabel-gif/Cadence★ 29same name
- shinyobjectz/cadence★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository8 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .cursor/rules/manifest-guardrails.mdcCursor rule
- llms.txtllms.txt
- capture-feedbackSkill
- context-hygieneSkill
- new-executionSkill
- new-motionSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 4 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
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 27
Read what is already in flight: Every initiatives/*.md with status: active.
- Very specific: line 50
If the week's work does not ladder up to the one goal in plan/company-plan.md, say so plainly instead of rationalizing it.
- Distinctive: line 76
First occurrence?: Leave it in the log. One observation is an anecdote.
- Very specific: line 91
Summarize: Write outputs/YYYY-MM-DD-weekly-review/README.md with:
- Distinctive: line 115
If nothing meets the repetition bar, say so.
- Distinctive: line 121
Append-only: Never edit yesterday's log or last week's outputs entry to make it look better.
- Distinctive: line 123
Do not fabricate: If the user says "good call today", ask which account rather than inventing an outcome.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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
The operating rhythm
cadence/ says what matters now. Three modes, one rhythm: plan on Monday, log
each day, review on Friday. Pick the mode from what the user asked for; if it is
ambiguous, ask rather than doing all three.
| Mode | When | Writes |
| Plan the week | Monday, or "what should this week focus on" | cadence/weekly/YYYY-Www.md |
| Log the day | End of day, or a recap of what happened | cadence/log/YYYY-MM-DD.md |
| Review the week | Friday, or "what happened recently and what did we learn" | outputs/YYYY-MM-DD-weekly-review/README.md plus context edits |
Mode: plan the week
The week is not a wish list. It is the slice of plan/ that gets done in the
next five days.
- Read the destination.
plan/company-plan.md(the one goal, the three moves) andplan/outcomes.md(what must become true, and what each outcome becomes ininfra/). - Read what is already in flight. Every
initiatives/*.mdwithstatus: active. Note their deadlines and unmet success criteria. - Read what was dropped.
cadence/carryover.md. For each item, bump the week count. If an item has now slipped three weeks or more, do not re-list it as a task. Surface it as a decision: either it gets an owner and a date this week, or it gets killed and moved to Resolved with the reason. An item that keeps slipping is a decision being avoided. - Read what just happened. The last week of
cadence/log/entries and any newoutputs/entries. What moved, what got stuck. - Write
cadence/weekly/YYYY-Www.mdfromcadence/weekly/_template.md:- One theme. A week with two themes is two weeks pretending to be one.
- The outcomes it serves, referencing
plan/outcomes.mdby id. A week that serves no outcome needs a very good explanation. - Deals to move, with the specific next move per account.
- Shipping: what leaves the building.
- Carried over, pulled from step 3.
- Report the plan back in chat: theme, outcomes served, and any item you escalated from carryover into a decision.
If the week's work does not ladder up to the one goal in
plan/company-plan.md, say so plainly instead of rationalizing it.
Mode: log the day
Two minutes, not a status report. The log exists so that next Monday's plan and next Friday's review are written from what actually happened rather than from memory.
- Write
cadence/log/YYYY-MM-DD.mdfromcadence/log/_template.md:- What moved. Concrete: a call happened, a play shipped, a number changed. Not "worked on outbound".
- What is stuck. Anything that did not resolve today.
- Worth remembering. A thing you heard that might become knowledge: an objection, a signal, a competitor mention, a reason a deal was won.
- Route the stuck items. Anything under "What is stuck" that will not
resolve tomorrow goes into
cadence/carryover.mdwith today's date and a week count of 1. Do not let it live only in the log, where it will be forgotten. - Route the observations. For each item under "Worth remembering", check
context/for an existing file on the same claim:- Second independent occurrence? It clears the repetition bar. Use the
capture-feedbackskill to promote it intocontext/with the rightconfidence:level. - First occurrence? Leave it in the log. One observation is an anecdote. Say so, and do not write it into context.
- Second independent occurrence? It clears the repetition bar. Use the
Mode: review the week
outputs/ is what happened. context/ is what we know. This mode moves
learning from the first to the second, so the repo gets smarter every week.
- Collect. List
outputs/entries from the last 7 days (directory names start with the date). Read each entry's summary. Then find every entry, whatever its date, still atoutcome: pending: an execution that merged and has finished gets its outcome filled from its batch or run (thenew-executionskill, step 7). One still pending after its merge was held; list it under open questions with the reason. - Summarize. Write
outputs/YYYY-MM-DD-weekly-review/README.mdwith:- What ran (plays, campaigns, research) and headline numbers.
- What worked, what failed, with evidence.
- For each outcome an agent owns in
plan/outcomes.md: the weekly signal against the pace the target needs, any guardrail breached, tests started and ended inside the experiment budget with their result against the bar they set in advance, proposals waiting for the owner, and whether the escalation point was reached. - Open questions.
- Extract durable learning. For each finding that changes what we know,
propose the matching context edit:
- A new objection heard twice or more:
context/objection/<slug>.md. - A win or loss with a clear reason: update
context/client/andcontext/proof/. - A signal that predicted a conversion:
context/signal/<slug>.md. - Messaging that outperformed: note it in the relevant
context/motion/. Apply the repetition rule: a claim goes into context after it shows up independently at least twice. Once is an anecdote.
- A new objection heard twice or more:
- Ship. One PR: the review entry in
outputs/plus the proposed context edits. The PR description separates "observed" (outputs) from "concluded" (context changes) so reviewers can push back on conclusions.
If nothing meets the repetition bar, say so; an empty context diff is a valid review.
Rules (all three modes)
- Absolute dates and ISO week numbers (
2026-W29), never "this week". - Append-only. Never edit yesterday's log or last week's outputs entry to make it look better; add a dated addendum instead.
- Do not fabricate. If the user says "good call today", ask which account rather than inventing an outcome. If the CRM is not readable from here, ask rather than invent deals or numbers.
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.

