ruflo-iot-cognitum
claudeIoT device lifecycle, telemetry anomaly detection, fleet management, and witness chain verification for Cognitum Seed hardware
5 skills · 1 command · 4 agents
Universal autonomous engineering loop (skills, agents, hooks) with cross-harness orchestrators.
Plugin dev-loop from marketplace dev-loop-marketplace in mios-dev/-dev-loop. Commands read from the repo's manifest — 1 month ago.
This plugin runs code on your machine — its hooks fire automatically on events like tool use and it starts local MCP servers. The exact commands are listed below; read them before installing.
backlogWork the task ledger that carries a repository between sessions - list what is unblocked right now, add a task with its acceptance criterion and both controls written before the code, flip a status with cited evidence, validate the file, re-render TASKS.md, and emit a lane object for /verify or /dev
dev-loop-webRuns a scheduled, out-of-loop build-and-verify task for an agent that has web access but no shell or repository checkout. Fetches the task's contract file fresh at a pinned commit, writes a definition of done, builds the deliverables in its own workspace, proves every check both ways with a planted
dev-loopAutonomous engineering loop and multi-harness worker orchestrator for ANY codebase, language, or build system. Use for every coding task that takes more than one iteration - bug fixes, red-to-green CI repair, refactors, migrations, feature work, dependency bumps, test-suite repair, security hardenin
gate-auditAudit a repository's STANDING gates - a CI workflow, a Makefile/justfile recipe, a lint or gate script, a test directory - against the SKILL.md section 7 "Checks That Cannot Fail" taxonomy, then prove the survivors. `gate_audit.py` detects seven mechanically decidable rows (Skip-as-Pass, Swallowed F
goalDefine an engineering goal with explicit stopping conditions, decompose it into tasks, and drive it to done through the dev loop (`/goal dev OBJECTIVE`), or evaluate whether the current goal's stopping conditions hold (`/goal eval`). Use when the user names an outcome to reach rather than a single c
handoffEnd a session on disk instead of in prose - flip the task with evidence, park uncommitted work as a patch, and append a ledger entry pinned to HEAD that states what is still unverified. Wraps `artifacts.py tasks set/render` and `adapters.py ledger` so the next session starts from files rather than f
initBootstrap a fresh MiOS development environment and start a dev-loop cycle in it - locate or shallow-clone the sibling repos (MiOS, mios-bootstrap, -dev-loop, mios-micro; the MiOS checkout is pinned to its GitHub origin), install what MiOS's one .devcontainer/Containerfile installs (Fedora host - its
postmortemDiagnose a lane that did not finish, from its run directory - separate a denied/empty harness turn, a lost worker with no receipt, an ownership stray, a vacuous gate, and a genuine defect, because each has a different remedy and only the last one is a code bug. Reads the receipts already on disk; ru
preflightCheck the environment before a run — which harness binaries exist, whether the exact CLI flags the lane command templates rely on are still in their --help output, Antigravity authentication and grants when agy is installed, and repo conformance when the cwd is a dev-loop checkout. Use before dispat
researchUpstream research before deciding - vendor docs, changelogs, release diffs (`base_ref..upstream_ref`), breaking-change audits, and technology evaluations, written into SPIKE / UPSTREAM_AUDIT / TECH_EVAL documents. Use before an architectural decision, a dependency bump, a migration, or when a deprec
reviewSCOPE staged code oversight of a diff (Staged Code Oversight with Proportional Escalation) - Stage 1 agent review (contract preservation, invariant and security audit, suppression/threshold drift, dead code, test strength), Stage 2 steering-developer ownership check, Stage 3 proportional peer escala
shipSafe trunk-based shipping of a lane or feature branch - pre-flight gates, --no-ff merge with abort-on-conflict, worktree and branch cleanup, CHANGELOG entry, optional tag. Use only after /review passed; ship is user-invoked, never auto-triggered.
teamLaunch full-teamwork multi-agent workflows where every harness participates - Claude Code, Antigravity/AGY, Codex, Gemini, Copilot, Cursor, OpenCode, and ANY OpenAI-compatible agent - coordinated through the dev-loop lane discipline and the loop-translation bridge. Distills the AGY teamwork protocol
triagePhantom-failure and flakiness triage for a failing command - separates deterministic bugs from flaky tests and environment/path/cache artefacts by repeated isolated runs, produces a minimal repro script and a TRIAGE_INCIDENT record. Use when a test or build fails, especially "passes locally, fails i
verifyRun the two-sided gate on work in progress - the positive control must pass, a planted negative control must FAIL naming the plant, and the tree must be restored unchanged. Wraps `adapters.py gate` so single-threaded work in any harness can reach the same gate an orchestrated lane gets, instead of a
watchReport the state of a dev-loop run from its run directory — per-lane model report, the shell-written exit receipt, job liveness and log tails — and name the anomalies (lost jobs, missing receipts, lanes with no report). Read-only; it never dispatches and never merges. Use when a multi-lane run is in
websearchTargeted web discovery for canonical vendor documentation, API specifications, changelogs, and exact error signatures. Use for a quick fact check during work; use /research for a full upstream audit.
auditorRead-only reviewer for the dev loop - SCOPE diff audit, suppression and threshold drift, secrets, verification-evidence check; also audits surviving lanes when the auditor lane died. Never edits files.
lane-workerL2 lane worker - implements exactly one lane contract inside an isolated git worktree, runs both controls, and returns a devloop_report. Dispatched by the orchestrator; not for direct use on unscoped tasks.
orchestratorL0 host for the dev loop - decomposes an objective into lanes with exclusive owned_paths, dispatches lane-worker subagents (worktree-isolated) or the cross-harness orchestrator, runs the two-sided merge gates, merges, closes tasks, writes the ledger. Use for parallel or multi-task objectives.
researcherRead-only upstream researcher - vendor docs, changelogs, installed package source, release diffs. Cites versions and URLs; never changes code.
triagePhantom-failure and flakiness diagnostician - repeated isolated runs, environment/path/cache checks, minimal repro. Diagnoses; does not fix.
dev-loopruns: python3 ${CLAUDE_PLUGIN_ROOT}/skills/dev-loop/scripts/devloop_mcp.py
"${CLAUDE_PLUGIN_ROOT}/hooks/session-start.sh"
"${CLAUDE_PLUGIN_ROOT}/hooks/prompt-context.sh"
"${CLAUDE_PLUGIN_ROOT}/hooks/owner.sh"
"${CLAUDE_PLUGIN_ROOT}/hooks/guard.sh"
"${CLAUDE_PLUGIN_ROOT}/hooks/no-env.sh"
"${CLAUDE_PLUGIN_ROOT}/hooks/no-merge.sh"
"${CLAUDE_PLUGIN_ROOT}/hooks/format.sh"
"${CLAUDE_PLUGIN_ROOT}/hooks/pre-compact.sh"
"${CLAUDE_PLUGIN_ROOT}/hooks/stop-gate.sh"
"${CLAUDE_PLUGIN_ROOT}/hooks/stop-gate.sh"
IoT device lifecycle, telemetry anomaly detection, fleet management, and witness chain verification for Cognitum Seed hardware
5 skills · 1 command · 4 agents
Find and verify professional email addresses, search contacts by domain, and enrich company data -- directly in Claude.
13 skills · 1 MCP server
Portable operating method: session lifecycle agents, pre-ship code review, deploy with live verification and state doc update, and production log health check
5 agents
45 advanced engineering skills: agent designer, agent workflow designer, RAG architect, database designer + schema designer + SQL assistant, migration architect, observability designer, dependency auditor, changelog generator, API design reviewer, API test suite builder, CI/CD pipeline builder, MCP server builder, skill security auditor, skill tester, performance profiler, focused-fix, browser-automation, full-page-screenshot, git-worktree-manager, monorepo-navigator, codebase-onboarding, interview-system-designer, runbook-generator, spec-driven-workflow, secrets-vault-manager, env-secrets-manager, pr-review-expert, self-eval, tc-tracker, feature-flags-architect, kubernetes-operator, chaos-engineering, ship-gate, slo-architect (error-budget + multi-window burn-rate alerts per Google SRE Workbook), tech-debt-tracker, agent-regression-evals, coding-agent-rollout, commit-craft, agent-run-forensics and i18n-localization. Agent skill and plugin for Claude Code, Codex, Gemini CLI, Cursor, OpenClaw.
48 skills · hooks
Persistent project planning files with lifecycle hooks that inject selected plan context. Automatic recovery uses project files only. Explicit catchup modes may read same-project local session records to emit aggregate counts or bounded nonce-framed excerpts. Optional gated mode can request continuation on capable hosts and never runs commands declared in Markdown. No network upload path. Installs across 60+ agents.
6 skills · 13 commands · hooks
PRD generation and conversion skills for the Ralph autonomous agent loop
2 skills
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.