weekly-retro
Skill4 starsChanged 4 months agoMITView on GitHub
Friday-afternoon blueprint that compiles the week's heartbeat-standups into one org retro — shipped, slipped, budget burn vs caps, the top lesson, and exactly one process change proposed to the human. Use when the scheduled run fires ("Run the org weekly retro"), or when the CEO or human asks "how did the week go", "run the retro", or "what should we change". Don't use for the daily per-agent check-in — that's heartbeat-standup.
7 other projects use a version of this.
Adopted by
Most recently taken up by
- 0xUrsanomics/utopia-ostheir own version, first seen Nov 1, 2026
- JamsusMaximus/growth-agentstheir own version, first seen Oct 25, 2026
- jp-solumhealth/jpstacktheir own version, first seen Oct 14, 2026
- serejaris/ris-claude-codetheir own version, first seen Oct 9, 2026
- serejaris/personal-corp-ostheir own version, first seen Aug 6, 2026
- DScardini91/agentic-ostheir own version, first seen Jun 26, 2026
Used with
Projects with it also use idea and seo-audit.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
szweibelSkillidea
Generate a wide field of options when the user is brainstorming, exploring, expanding, or pressure-testing — not recommending or implementing
Also in 3 projects that use this
AnthropicSkillseo-audit
Run a comprehensive SEO audit — keyword research, on-page analysis, content gaps, technical checks, and competitor comparison. Use when assessing a site's SEO health, when finding keyword opportunities and content gaps competitors own, or when you need a prioritized action plan split into quick wins and strategic investments.
Also in 3 projects that use this
Other versions7 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 7
- serejaris/personal-corp-os★ 244same name
- serejaris/ris-claude-code★ 244same name
- JamsusMaximus/growth-agents★ 21same name
- Joshua-Palamuttam/workbench-template★ 3same name
- 0xUrsanomics/utopia-os★ 2same name
- DScardini91/agentic-os★ 2same name
- jp-solumhealth/jpstack★ 0same 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 repository74 other files it gives its agents.
- agent-eval-harnessSkill
- context-budget-managerSkill
- incident-rollbackSkill
- memory-hygieneSkill
- model-cost-optimizerSkill
- prompt-debuggerSkill
- skill-author-coachSkill
- third-party-skill-vetterSkill
- api-endpoint-testerSkill
- bug-triage-protocolSkill
- changelog-release-notesSkill
- docs-readme-writerSkill
- env-secrets-hygieneSkill
- github-pr-reviewSkill
- git-ship-flowSkill
- lovable-brief-writerSkill
- supabase-migration-writerSkill
- vercel-deploy-checkSkill
- cold-outreach-sequencerSkill
- competitor-pricing-scanSkill
- contract-redline-scanSkill
- customer-support-inboxSkill
- invoice-runnerSkill
- kpi-dashboard-weeklySkill
- lead-research-dossierSkill
- payment-reconciliationSkill
- pipeline-trackerSkill
- proposal-builderSkill
- stripe-payment-linkSkill
- testimonial-harvesterSkill
- weekly-revenue-reportSkill
- brand-voice-guardianSkill
- case-study-builderSkill
- content-repurposerSkill
- deep-research-briefSkill
- email-nurture-sequenceSkill
- landing-page-copywriterSkill
- linkedin-content-packSkill
- newsletter-assemblerSkill
- seo-blog-writerSkill
- social-listening-reportSkill
- x-thread-builderSkill
- youtube-script-writerSkill
- budget-guardrailsSkill
- ceo-operating-doctrineSkill
- delegation-protocolSkill
- editor-personaSkill
- heartbeat-standupSkill
- hiring-briefSkill
- ops-agent-personaSkill
- researcher-personaSkill
- sales-agent-personaSkill
- support-agent-personaSkill
- writer-personaSkill
- calendar-conciergeSkill
- decision-journalSkill
- email-triageSkill
- file-organizerSkill
- habit-trackerSkill
- inbox-zero-sprintSkill
- The first 60 shown.
The file
Files
Weekly Retro
Runs Friday at 16:00 and turns a week of heartbeat-standups into one honest page: what shipped (with ticket links), what slipped (with owners and reasons), where the money went against the caps, the single most valuable lesson, and one — exactly one — process change proposed to the human. The retro reports and proposes; the human decides. A retro that flatters the org is worse than no retro.
When to Use
- The scheduled Friday run ("Run the org weekly retro").
- The CEO or human asks for a week-in-review, budget week summary, or "what should we change".
- A rough week ends and the org needs the slippage laid out plainly.
- Not for: the daily per-agent check-in (
heartbeat-standup— this skill consumes those), mid-cycle budget alarms (budget-guardrailsfires immediately, not on Fridays), or enacting the process change (proposal only; human approves, CEO implements).
Role Boundary
- This run owns: collecting the week's standups, verifying claims against tickets, computing budget burn vs caps, distilling the top lesson, and drafting the one process-change proposal.
- It escalates: the whole deliverable (to the human — the retro's entire job is escalation with evidence), agents with missing standups (named, to CEO), budget breaches discovered during compilation (to CEO +
budget-guardrailsimmediately, not held for the report), and any integrity gap between standup claims and ticket states. - It never: implements process changes, edits or backfills agents' standups, marks unverified work as shipped, or softens numbers to keep the mood up.
Quick Reference
| Action | Rule |
| Inputs | All heartbeat-standups since last retro + ticket states + budget ledger |
| Shipped | Only ticket-verified completions with goal ancestry; "basically done" = slipped |
| Slipped | Every miss gets owner + reason + new date or explicit descope — no silent drops |
| Budget | Per-agent and org burn vs caps; >80% flagged, breaches escalated on discovery |
| Top lesson | One, chosen for expected impact next week, evidence-linked |
| Process change | Exactly one proposal, smallest-change-that-helps, human decides |
Procedure
- Collect — pull every heartbeat-standup since the last retro, the ticket board state, and the budget ledger. Missing standups don't stop the run: name the agent and the gap in the report ("no standup Wed–Thu: {agent}") and compile from tickets for them.
- Verify shipped — for each claimed completion, check the ticket: marked done, deliverable attached, goal ancestry intact (task → project → goal → mission). Verified → Shipped with ticket link and the goal it advanced. Claimed-but-unverifiable → Slipped, with the discrepancy noted honestly.
- Compile slipped — everything planned-not-done: owner, one-line reason (real reason, not weather), and disposition — new date, descoped with CEO note, or blocked-on-human with the exact blocker. A slip without a disposition is not allowed to leave the retro.
- Compute budget burn — per agent and org-wide: spent vs cap, week-over-week trend. Anything >80% of cap gets flagged; an outright breach found during compilation goes to the CEO and
budget-guardrailsthe moment it's found — the Friday report is not an excuse to sit on a breach for three days. - Distill the top lesson — scan standups and slips for the pattern that cost or taught the most. One lesson, stated as a falsifiable observation with linked evidence ("3 of 4 slips traced to briefs missing audience — tickets #41, #44, #47"), not a platitude.
- Propose one process change — the smallest concrete change that addresses the top lesson or the week's dominant friction: what changes, who's affected, expected effect, how next retro will measure it, and rollback if it doesn't help. One proposal only — a list of five is a way of deciding nothing.
- Deliver — post the Output Template to the human and the org channel. File the retro where previous retros live, and check last week's process change (if one was approved): did it measurably help? Report that verdict too.
Output Template
# Weekly Retro — week ending {date}
**Standups received:** {n/expected} (missing: {agents or "none"})
## Shipped ({n})
- {deliverable} — {agent} — ticket #{id} — advanced: {project → goal}
## Slipped ({n})
- {item} — {owner} — why: {reason} — disposition: {new date / descoped / blocked-on-human: {blocker}}
## Budget
| Agent | Spent | Cap | % | Trend |
|---|---|---|---|---|
Org total: {x/y} — flags: {>80% list, breaches already escalated}
## Top lesson
{one falsifiable observation} — evidence: {ticket/standup links}
## Last week's change: {change or "none approved"} — verdict: {helped / didn't / too early, with the measure}
## Proposed process change (one)
**Change:** ... | **Affects:** ... | **Expected effect:** ... | **Measured by:** {next-retro check} | **Rollback:** ...
→ Awaiting human decision.Pitfalls
- Optimism laundering — "90% done" work lands in Shipped and the human plans on phantom progress. Recovery: Shipped is a ticket state, not a feeling; re-run step 2 and move anything unverified to Slipped with the honest note.
- Slips without dispositions — the same item slips three weeks running because it's reported and never re-planned. Recovery: every slip leaves the retro with a new date, a descope, or a named blocker; check the previous retro's slips first and flag repeat offenders explicitly.
- Process-change buffet — five proposals ship, the human rubber-stamps none, nothing changes. Recovery: rank by expected impact, keep #1, park the rest in a backlog note; one decidable proposal beats five ignorable ones.
- Held breach — a budget breach found Tuesday-in-the-data waits for the Friday report. Recovery: escalation is on-discovery per step 4; if it was held, escalate now and say in the retro that it was found late — the ledger of trust matters more than a clean-looking process.
- Lesson platitudes — "we should communicate better" as the top lesson, teaching nothing. Recovery: force the falsifiable form — what pattern, in which tickets, costing what; if no evidence links, it's not the lesson, keep looking.
Verification
- Every Shipped item has a done-state ticket link and its goal ancestry named
- Every Slipped item has owner, real reason, and a disposition
- Budget table covers all agents plus org total; >80% flagged; any breach escalated on discovery, not held
- Missing standups named, not papered over
- Exactly one process change proposed, with measure and rollback, marked awaiting human decision
- Last week's approved change (if any) got a measured verdict
- Nothing was implemented, edited, or backfilled by the retro run itself
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.

