REUSABLE METHODS FOR AI AGENTS

Keep the method your agent discovered.

Rote turns a successful agent run into an inspectable, repeatable Play that can travel across harnesses, models, machines, and teams.

START WITH PLAY
curl -fsSL https://getrote.dev/playoffs/install.sh | sh

Installs Play and Rote, detects supported harnesses, and starts guided setup.

  1. 01INSTALL PLAY
  2. 02GUIDE ONE RUN
  3. 03GET A REUSABLE PLAY
≡×AGENTEXPLORING

USER GOALIS MY WEBSITE READY TO LAUNCH?

  1. 01play searchNO MATCH
  2. 02API callOK
  3. 03browser checkFAILED
  4. 04retry + correctionOK
≡×ROTE · TRACERECORDING

EACH COMMAND BECOMES A CONNECTED TRACE SPAN

✓ VERIFIED{ } PARAMETERIZED7/7 TESTED
≡×PLAYREADY
PLAY CREATEDWEBSITE LAUNCH READINESShttps://play.modiqo.ai/modiqo/[email protected]

v1.1.2INSPECTABLE

  1. ✓URI MINTEDSIGNED + VERSIONED
  2. ✓READY TO ROUTEPRIVATE OR COMMUNITY

ONE PLAY. ANY SUPPORTED AGENT.

Claude CodeHermes AgentCodex AppKimiCursorPi Coding Agent
WHAT IS A PLAY?

A successful trace, made executable.

A Play runs the method again with fresh inputs instead of re-exploring the task.

WHAT A PLAY CONTAINS

A Play is inspectable by design.

WHERE THE URI CAN GO

Move proven work beyond the frontier model.

Spend frontier intelligence discovering the method. Move repeatable runs to lower-cost models. Share the URI so others can run, inspect, and improve it.

TRANSFER RECORDPLAY · MODIQO/HELLO
01DISCOVERED BYFRONTIER AGENTREASON ONCE
PLAY URIhttps://play.modiqo.ai/modiqo/hello
02ALOWER-COST MODELRECALL
02BTEAMRUN + REVIEW
02CCOMMUNITYINSPECT + IMPROVE
ONE URIFRESH INPUTSLOCAL CREDENTIALS
WHEN REALITY CHANGES

Recall stops when reality changes.

When the Play's assumptions no longer hold, the agent stops following it and reasons again. The successful new path can become the next version.

EXECUTION GATEPLAY MATCH · VERSION 1
01ASSUMPTION CHECKVERIFY CURRENT CONDITIONS
UNCHANGEDRECALL V1RUN KNOWN PATH
CHANGEDREASON AGAINSUCCESSFUL PATH → V2
THE PLAY DOES NOT OVERRIDE REALITY
PUBLISHED PLAYS

Real Plays. Visible evidence.

Inspect the publisher, version, permissions, tools, and digest before an agent runs the Play.

PUBLISHED CONTRACTVERSIONAUTHOR / PUBLISHEREXECUTION + GUARDS
Website launch readiness↗

Checks DNS, TLS, headers, metadata, and Lighthouse.

VERSIONv1.1.2
AUTHOR / PUBLISHERChetanPUBLISHED BY MODIQO
PYTHON · DNS · LIGHTHOUSE
8 STEPSREAD ONLYDIGEST VERIFIED
Summarize stacked GitHub PRs↗

Builds and explains a complete GitHub pull-request stack.

VERSIONv0.2.1
AUTHOR / PUBLISHERDebasish GhoshPUBLISHED BY MODIQO
GITHUB · SHELL
3 STEPSREAD ONLYDIGEST VERIFIED
Retrieve rideshare receipts↗

Finds, deduplicates, and totals rideshare receipts.

VERSIONv0.1.6
AUTHOR / PUBLISHERCode BotPUBLISHED BY MODIQO
GMAIL · LOCAL CREDENTIALS · CONSENT
4 STEPSREAD ONLYDIGEST VERIFIED
MEASURED ON MODIQO/HELLORepeat the work, not the reasoning.
DISCOVER14,900
REPEAT~300
98%LESS REASONING
ONE LIBRARY · EVERY TEAM

Build a Play library for every team.

Sales, Marketing, DevOps, and AgentOps keep proven methods in one shared library. Each team keeps its preferred harness and local credentials.

FROM SUCCESSFUL JOB → SHARED PLAY4 TEAMS · 4 JOBS · ONE LIBRARY
CHOOSE A TEAM
SALES · SUCCESSFUL AGENT JOB
TEAM-SPECIFIC WORKsales, launches, incidents, skillsONE SHARED LIBRARYproven methods stay discoverableANY HARNESSteams keep the tools they preferLOCAL CREDENTIALSaccess never travels with the Play
YOUR NEXT PLAY

What should your agent never solve twice?