playwright
01MCP server Claude CodeCodexCursor +2
MCP server "playwright", hosted remotely at localhost, as configured in kesteva/cyboflow.
not rated 65
+2 8d ago A
tokens not measured
original MIT
Agent Claude Code
Compound load subagent. Surveys the recently merged/completed work the orchestrator names — git log/diff against the base branch plus the run-context digest — and returns ONE Merged work summary. Surveys only: it never mines learnings, never judges, and never writes cyboflow state.
not rated 65
+2 8d ago A 0 tokens
original MIT
Agent Claude Code
Compound write-back subagent. Applies the learnings the human approved — quick fixes and instruction-file edits — in place in the worktree, and reports back which landed. Edits files only: it never re-decides what was approved and never writes cyboflow state.
not rated 65
+2
changed 1mo ago A 0 tokens
original MIT
Agent Claude Code
Compound extract subagent. Mines the Merged work summary the load step produced (plus the diff and run-context digest) for durable learnings that clear an explicit recurrence/impact bar, each tagged quick / task / doc, plus a discarded list, with evidence. Returns the draft learnings; never writes cyboflow state.
not rated 65
+2
changed 1mo ago A 75 tokens
original MIT
Agent Claude Code
Planner adversarial-design reviewer (optional). Read-only critic that stress-tests the idea spec and any UI prototype or architecture, returning concrete must-fix defects and advisory findings. Never writes cyboflow state.
not rated 65
+2
changed 26d ago A 51 tokens
original MIT
Agent Claude Code
Planner architecture-design subagent (optional). Proposes a human-reviewable architecture for a structurally complex idea and returns a markdown section for the orchestrator to fold into the idea body. Never writes cyboflow state.
not rated 65
+2
changed 1mo ago A 52 tokens
original MIT
Agent Claude Code
Planner context-gatherer. Probes user intent, produces a short idea stub for approval, then expands the approved immutable stub into a full idea spec. Read-only — returns content for the orchestrator to persist; never writes cyboflow state.
not rated 65
+2 8d ago A 57 tokens
original MIT
Agent Claude Code
Planner epics subagent (large ideas only). Decomposes an idea into epics with dependency edges and returns the proposed breakdown for the orchestrator to persist. Never writes cyboflow state.
not rated 65
+2
changed 21d ago A 49 tokens
original MIT
Agent Claude Code
Launch interview subagent. Drives an in-depth multi-round project interview, synthesizes the approved answers into a project brief, then decomposes the brief into an ordered, dependency-sequenced idea set. Read-only — returns content for the orchestrator to persist; never writes cyboflow state.
not rated 65
+2
changed 1mo ago A 67 tokens
original MIT
Agent Claude Code
Planner research subagent (optional). Pulls external docs, prior art, and library/API references relevant to an idea and returns notes for the orchestrator to fold into the idea body. Never writes cyboflow state.
not rated 65
+2 8d ago A 52 tokens
original MIT
Agent Claude Code
Planner tasks subagent. Breaks an idea (or each epic) into concrete, independently shippable tasks with acceptance criteria, and returns them for the orchestrator to persist. Never writes cyboflow state.
not rated 65
+2
changed 1mo ago A 51 tokens
original MIT
Agent Claude Code
Planner UI-prototype subagent (optional). Builds a self-contained static HTML+CSS mockup of the idea's UI and writes it to the run artifacts dir for the orchestrator to report as the ui-prototype artifact. Never writes cyboflow state.
not rated 65
+2
changed 1mo ago A 61 tokens
original MIT
Agent Claude Code
Sprint address-review subagent. Takes the code-review findings this run filed, independently verifies each against the actual code, judges which are worth acting on now, and fixes those in place — returning a per-finding disposition (FIXED / DEFERRED / INVALID) the orchestrator uses to resolve or keep each one. Runs…
not rated 65
+2 8d ago A 83 tokens
original MIT
Agent Claude Code
Sprint code-review subagent. Inline review of the task diff for correctness and naming/layering/pattern compliance. Classifies by severity — in-scope must-fix defects go in ## Blocking (the orchestrator loops the implementer back to fix them), out-of-scope issues go in ## Findings (filed for human triage). Never…
not rated 65
+2
changed 23d ago A 88 tokens
original MIT
Agent Claude Code
Sprint dependency-analyzer subagent. Reads the sprint's seeded tasks (bodies, acceptance criteria, expected files) and proposes task→task blocking edges so the orchestrator can run independent tasks in parallel and dependents in order. Returns proposed edges with reasons; never writes cyboflow state.
not rated 65
+2
changed 1mo ago A 69 tokens
original MIT
Agent Claude Code
Sprint implement subagent. Writes the diff for a single ready task, scoped to its acceptance criteria, and runs local checks. Returns a summary of the change; never writes cyboflow state.
not rated 65
+2
changed 23d ago A 47 tokens
original MIT
Agent Claude Code
Runbook-bootstrap subagent. Deployed by the main-process controller when a sprint/ship lane's visual verification would be skipped for want of a verification runbook. Surveys the project READ-ONLY and returns a portable runbook for one modality — or an honest "not possible" — as structured data. Writes nothing: no…
not rated 65
+2
changed 19d ago A 0 tokens
original MIT
Agent Claude Code
Sprint review subagent. Cross-task taste pass over the WHOLE sprint's combined diff — coherence between lanes, duplicated helpers, seam mismatches, CLAUDE.md drift — returning issues as findings for the orchestrator to record. Never writes cyboflow state.
not rated 65
+2
changed 1mo ago A 62 tokens
original MIT
Agent Claude Code
Sprint full-suite verifier. Runs the project's full test suite once after the task work is complete and returns a PASS/FAIL verdict. Never writes cyboflow state.
not rated 65
+2 8d ago A 43 tokens
original MIT
Agent Claude Code
Sprint task-verify subagent. Checks the diff against the task's acceptance criteria and returns a PASS/FAIL verdict with per-criterion evidence for the orchestrator's loopback decision. On PASS it also composes the visual-verification task (or declares one not applicable) for the central visual verifier. Never writes…
not rated 65
+2
changed 19d ago A 77 tokens
original MIT
Agent Claude Code
Central visual-verification agent. Deployed per request by the main-process verification scheduler in an isolated snapshot worktree; builds and serves the deliverable, drives the composed behaviors, captures screenshots, and returns a structured verification report. Never writes cyboflow state.
not rated 65
+2
changed 18d ago A 62 tokens
original MIT
Agent Claude Code
Sprint test-author subagent. Adds unit/integration tests covering the new diff and runs them; bootstraps a minimal runner when the project has none, or skips loudly with a machine-readable outcome. Never writes cyboflow state.
not rated 65
+2
changed 23d ago A 55 tokens
original MIT
Agent Claude Code
Verify Setup subagent. Surveys a project for how its UI actually stands up (scripts, framework, Electron vs web, isolation levers, existing runbook), then drafts a portable verification runbook per modality with a required attestation channel, machine-local bindings, and the lowest-rung repo changes that make it work.…
not rated 65
+2
changed 19d ago A 104 tokens
original MIT
Plugin Claude Code
Plugin marketplace listing 2 plugins: soloflow, soloflow-dev.
not rated 40 4mo ago A
tokens not measured
original MIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: