claudecodex

frends

Work with your Frends tenant through the Frends Platform MCP: find and inspect Processes, diagnose failures, plan integrations, build and review drafts, and run Processes, and nothing leaves the draft stage without your decision.

1 on GitHubby Frendsv0.10.0MITupdated 13 days agoproductivity
/

Plugin frends from marketplace frends in Frends-Product-Management/frends-claude-plugin. Commands read from the repo's manifest — 13 days ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (13)

  • agent-contract

    What the frends agents work under: done as evidence, fix caps, what is never written, the draft boundary and criteria. Preloaded into the agents; not a task in itself.

  • build-a-process

    Build or edit a Frends integration Process as a draft, up to a passing validation. Use when someone wants to create, scaffold, draft or change an integration, or add a trigger or a Task to a Process. Do not use to run a Process, to review a finished draft, to choose between Process shapes that are s

  • build-loop

    Run a bounded build loop that takes an integration request or a confirmed plan to a validated Frends Process draft, with a run record under .frends/, an independent review before success, and a named terminal state. Use when a plan carries acceptance criteria, when several build-and-validate rounds

  • deliver-loop

    Run the delivery loop over a confirmed integration plan, building each planned Frends Process to a validated draft with two independent reviews per Process and a handoff at the end. Invoked by name only; a plan whose confirmation status is not confirmed is turned away at the door.

  • diagnose-failures

    Investigate failed, stuck or slow runs of a deployed Frends Process and explain the cause, read-only. Use when someone reports that an integration is failing, erroring, stuck or not producing results, or asks what went wrong with a run. Do not use for a connection that does not work at all, such as

  • find-and-inspect

    Find Processes in a Frends tenant and explain how one is built, without changing anything. Use when someone asks what integrations exist, where a Process is deployed, which version is live, what a Process does internally, whether something that already does a job exists, or which Tasks and environme

  • fix-loop

    Run a bounded fix loop for a Frends Process that is failing in an Environment, from diagnosis through the person's chosen cause to a validated edit draft, with a run record under .frends/ and a named terminal state. Use when a deployed Process fails and the person wants the fix built, not only expla

  • getting-connected

    Connect an AI client to a Frends tenant, and work out why the connection itself is not working. Use when the Frends tools are missing, when a call returns 401 or 404, when lists come back empty, when one specific tool is absent, when someone is setting up Frends for the first time, or when someone a

  • harness

    The rules a frends loop runs under: how a run ends, who reviews whom, the run record, and clients with no agents or hooks. Read at the start of every loop; not a task.

  • integration-planning

    Interview someone about an integration they want, one topic at a time, and turn their answers into a written integration plan the build step can work from, synthesize the same plan from a conversation that already happened, or write a questionnaire for a question only an absent colleague can answer.

  • process-patterns

    Choose the right shape for a Frends integration Process before building it, and say which shapes go where. Use when the requirements are known but the structure is not, when someone asks what kind of Process fits a need, or wants to compare approaches before any building starts. Do not use when the

  • review-a-draft

    Read a built Frends Process draft and report what is wrong with it, against the plan it came from and against Frends conventions, before anyone promotes it. Use when a draft is finished and someone asks whether it is right or complete, or wants it checked before promoting. Do not use to build or cha

  • run-a-process

    Run, start, trigger or test a deployed Frends Process, only on the person's explicit confirmation. Use when someone asks to run, execute, kick off, fire or test a Process, or to try out a draft they just built. Do not use to build or edit a draft, to review one, or to investigate a run that already

Subagents (3)

  • draft-reviewer

    Reads one Frends Process draft snapshot on one review axis, either Frends conventions or the integration plan, and returns findings only, in a fixed shape a gate checks. Dispatched by the review-a-draft skill and the loops; do not use to build, change, promote or run anything.

  • failure-diagnoser

    Reads the failing runs of one deployed Frends Process and returns three to five ranked causes with the run evidence behind each, read-only. Dispatched by the fix loop; changes nothing, proposes no fix, and never speculates past the evidence.

  • process-builder

    Builds or edits one Frends Process draft from a brief that carries frozen acceptance criteria and a pattern reference, slice by slice, up to a passing validate_process, and reports the last validation verbatim. Cannot promote, deploy, run, import a Task package or create an environment variable. Dis

MCP servers (1)

  • frends

    connects to: ${FRENDS_MCP_URL}

Hooks (4)

  • SessionStart

    node

  • PostToolUse

    node

  • Stop

    node

  • SubagentStop

    node

View source on GitHub

Similar plugins

caveman

claude

Talk like caveman. Cut filler. Keep technical details explicit.

22 skills · 1 command · 3 agents · hooks

JuliusBrussee/caveman 119,462

claude-mem

claudecodex

Persistent memory for Claude Code: the next session starts where the last one left off.

29 skills · 1 MCP server · hooks

thedotmack/claude-mem 116,618

Taste Skill

claudecodex

tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.

1 skill

Leonxlnx/taste-skill 114,391

career-ops

claudecodex

Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications

5 skills · hooks

career-ops-hq/career-ops 82,253

mem0

claudecodex

Cross-session memory and token savings for coding agents.

6 skills · 1 agent · 1 MCP server · hooks

mem0ai/mem0 68,893

pake

claudecodex

Package any website or local web build into a lightweight desktop app with Pake CLI.

1 skill

tw93/Pake 63,463

Maintain frends?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex pluginslisted