What did we actually ship?
Accountability through outcomes, not status updates.
Standups become recitals. People report what they touched, not what moved. Nobody connects daily work to quarterly goals.
By Friday, nobody remembers what actually shipped. Managers reconstruct the week from Slack threads and gut feel. Evidence evaporates.
Cross-team connections happen by accident. Two people solve the same problem in parallel. Leverage points go unnoticed until it's too late.
Every team has outcomes on a slide somewhere. The question is whether anyone can trace this week's work back to them — with evidence, not narrative.
An Amplifier bundle that automates the full weekly accountability cycle — from commitment logging to leadership showcases.
Each team member defines quarterly outcomes — observable, time-bounded, with short IDs. The anchor for everything that follows.
Weekly commitment logging ties this week's work to those outcomes. Interactive collection, file writing, cascade, and git commit — all automated.
Daily GitHub activity pull fetches real events — commits, PRs, reviews. Builds per-person activity files. Evidence, not self-reporting.
End-of-week: generate outcome-first audits, coordination analysis across teams, and narrative showcases for leadership. All from real data.
Meeting transcripts fetched via Graph API — processed into structured commitments
Commits, PRs, reviews pulled daily — evidence of what actually happened
Per-person weekly files tied to quarterly outcomes by ID
Quarterly goals with short IDs — the anchor for all accountability
11 declarative recipes orchestrate the entire cycle. Each writes its own files and commits to git independently.
Auto-process-standup runs M–Tu–W–Th at noon. Pull-activity runs daily. No human triggers needed.
Showcases use a staged recipe gate — manager reviews drafts before they post to Teams via webhooks.
The weekly-cycle recipe sequences three sub-recipes into a single end-of-week run. Each writes files and commits to git; the orchestrator sequences and reports.
Fetches GitHub events, builds per-person activity files, runs the github-activity cascade (Backlog items), commits everything in one shot.
Identifies the week's data files, processes per-person audit and review files in parallel, then assembles team-level views in a single final step.
Reads all weekly reviews and audit views. Synthesizes cross-team connection signals, themes, and leverage points.
The weekly cycle produces the complete accountability picture: what each person shipped, how it maps to outcomes, and where teams are unknowingly working on the same problems.
GitHub activity is pulled automatically. Standup transcripts are processed by the agent. Audits cross-reference commitments against actual work. You can't game what you didn't do.
The coordination recipe reads all weekly reviews across teams and synthesizes connection signals, shared themes, and leverage points — surfacing collaboration opportunities that no individual team can see.
Draft-showcase is a staged recipe: the agent generates outcome-first narratives, but a manager reviews and approves before anything posts to Teams via webhooks. Automation with judgment.
Every recipe writes files and commits to git. Commitments, activity, audits, coordination views — all versioned, all auditable, all diffable. No SaaS database. Just a repo.
Repository: cpark4x/amplifier-team-pulse (private)
Primary contributors: cpark4x, samueljklee
Bundle owner: kenotron-ms
Last updated: May 19, 2026
Data sources for this deck:
made-team-knowledge-data/capabilities/cpark4x/amplifier-team-pulse/team-pulse.yaml (type: bundle, owner: kenotron-ms)amplifier-team-pulse.yaml (type: tool, owner: cpark4x)auto-process-standup, confirm-alignment, download-meeting, draft-showcase, generate-coordination, log-commitments, prep-audit, process-standup, pull-activity, set-outcomes, weekly-cyclekenotron-ms/lifeos-vault capability: amplifier-team-pulse.yamlMethodology: All recipe names, descriptions, contributor lists, dependency chains, and automation schedules were extracted directly from team-knowledge YAML metadata. No runtime metrics were fabricated. The "4 auto-runs per week" figure reflects the auto-process-standup recipe description: "Runs at noon M-Tu-W-Th via agent-daemon." Recipe count (11) is the distinct recipe-type capabilities in team-knowledge. Contributor count (3) reflects the union of contributors across all capability files: kenotron-ms, cpark4x, samueljklee.
What this deck does not cover: Source line counts, commit history, and test coverage are unavailable — the repository is private and not cloned locally. All claims are grounded in team-knowledge metadata only.
Team Pulse replaces status theater with outcome-driven accountability. Set your quarterly goals, and let the system prove what you shipped.