timps-swarm

Implement ABHA enrollment, HIP/HIU flows, FHIR R4 bundles, and consent workflows (India ABDM).

Design A/B tests with Welch's t-test power analysis and GrowthBook configuration.

Audit a frontend (HTML / React / Vue) for WCAG 2.2 AA & AAA issues — colour contrast, ARIA, keyboard nav, focus order, screen reader hints.

Turn a technical decision into a Michael Nygard-style Architecture Decision Record: context, options matrix, decision, consequences, and a rendered Markdown file.

Generate a new fully-functional Python agent file from a natural-language description.

Curate the best AI agent / MCP / LLM-tool marketplaces and recommend picks by use-case. Tracks pulse.mcp.run, glama.ai, mcp.so, OpenAI GPT Store, Claude tool directory.

Forecast Indian agri-commodity prices (mandi arrivals, MSP, weather, festival demand). Returns direction, confidence, and a hedge recommendation for farmers / FPOs / traders.

Audit LLM applications against OWASP LLM Top 10 (LLM01-LLM10) with mitigations.

Turn a plain-English multi-step workflow into executable code for LangGraph, Temporal, or Claude-Flow — plus a Mermaid diagram and tests.

Set up PostHog/Mixpanel/Amplitude event tracking, funnels, and dashboards.

Diff two API specs (OpenAPI/AsyncAPI) and flag breaking changes, deprecations, and safe additions. Returns a semver recommendation and a migration guide for clients.

API Design Agent — generates a complete OpenAPI 3.1 spec (YAML) from a natural-language description. Enforces REST/GraphQL best practices, versioning strategy, and consistent naming BEFORE code generation starts. Returns the spec + a design rationale document.

Profile an API from access logs / traces — p50/p95/p99 by endpoint, slow query, N+1 detection, and concrete optimisations.

Run an OWASP API Security Top-10 (2023) audit against an OpenAPI spec or live endpoint — broken auth, BOLA, mass assignment, rate-limit, SSRF, injection — with a runnable test harness.

Audit backup health: Time Machine last backup time, all local git repos with uncommitted/unpushed changes, large files on Desktop/Documents not tracked by git or cloud. Returns a risk score and backup script.

Intelligently decompose a bulk task into parallel sub-tasks, execute them concurrently, and return a single aggregated result. Use this to do the same operation across many files at once. Examples: 'add tests for all handlers', 'review all API endpoints', 'add docstrings to every Python file', 'find flaky tests across all modules'.

Identify battery drainers. Reads battery %, health cycle count, and ranks processes by CPU usage (proxy for energy drain). Generates kill commands for the top energy vampires.

Scaffold a new project: folder structure, stub files, setup commands. Supports any language or framework — describe what you're building.

Generate Playwright E2E tests with page objects, fixtures, and CI config.

Re-design a noisy calendar: time-block deep work, batch meetings, protect maker time, add commute/exercise buffers, and emit an ICS patch and a focus-mode script.

Check TLS certificate expiry for all hosts found in config files. Warns 30 days before expiry and generates certbot renewal commands.

Generate a Conventional Commits changelog from commit history, PRs, and issues; supports semver bump recommendation.

Score accounts/users for churn risk from behavioural signals, build a survival-curve forecast, segment at-risk cohorts, and recommend save plays per segment.

Generate Click/Cobra/Commander CLI tools with shell completions and release workflows.

Scan Terraform configs for waste (over-provisioned instances, missing deletion protection) and check AWS CLI for idle resources.

Analyse git history to build an AST import graph, risk map, and tribal knowledge document.

Write a 5-7 step cold-outreach email sequence for a persona + offer, with subject A/B variants, social proof, objection handling, and a reply-promise CTA. Spam-safe.

Monitor competitors for pricing changes, new features, job postings, and public sentiment. Returns a competitive intelligence brief.

Map a codebase + infra config against SOC 2 / ISO 27001 / HIPAA / PCI-DSS / GDPR controls; produce an evidence pack and gap report.

Plan a GUI / computer-use task and emit OS-specific step scripts (cliclick / xdotool / PowerShell) with safety approval gates. Pairs with the runtime computer_use tools.

Auto-configure TIMPS Swarm as an MCP server in every AI coding tool installed on this machine (Cursor, Windsurf, GitHub Copilot, Cline, Continue, Aider, Goose, OpenCode, Gemini CLI, and more). Use --dry-run to preview what would change.

Scan a container image / Dockerfile for CVEs, misconfigurations, and supply-chain risk. Wraps trivy/grype/syft when present; falls back to static analysis of the Dockerfile.

Transform a long-form technical post into multiple short-form formats: tweet thread, LinkedIn post, changelog entry, docs blurb, and HN title.

Get a 3-sentence resumption briefing from the Context Keeper: what you were working on, current state (uncommitted changes, broken tests), and recommended next step. Use when the user says 'what was I doing' or 'catch me up'.

Analyse current work context: active apps, estimated browser tabs, current git branch and recent commits. Identifies distraction apps and generates a focus-mode script to quit them.

Review a contract (MSA, NDA, SaaS, vendor agreement) for red-flags: indemnification, liability caps, IP, data, termination, SLAs. Returns a clause-by-clause scorecard and a redline.

Estimate monthly/annual cloud cost for an architecture and generate a prioritised list of savings optimisations.

Summarise a court judgment / order: bench, facts, issues, holdings, reasoning, ratio, and citations in a structured brief. For legal research, not legal advice.

Design dbt/Airflow/Prefect/Dagster ETL/ELT pipelines with monitoring and lineage.

Validate and clean ML training datasets; generate data cards and quality reports.

Clean and normalise messy data: CSV, JSON, PDF extracts, copy-pasted tables. Returns cleaned records, a quality score, and SQL insert hint.

DB Agent — schema design, query optimisation, and migration scripts. Given a description of your data model it produces: CREATE TABLE SQL, an ER diagram in Mermaid, index recommendations, a migration script (Alembic / Flyway / raw SQL), and query templates. Pairs with the sql_injection LoRA for secure query generation.

Plan a zero-downtime database migration: expand-contract phases, backfill scripts, dual-write strategy, kill-switches, rollback plan.

Multi-hop, multi-source research planner: decomposes a query, runs parallel web searches, cross-references, and returns a sourced report with [N] citations and a bottom-line answer.

Delegate a multi-step goal to the TIMPS Agent Kernel. The kernel plans, routes to specialist agents, and returns a structured result. Use for complex goals like 'fix the auth bug and ensure 80% test coverage'.

Produce a SKU/location-level demand forecast with seasonality, promotions, and trend. Returns intervals, backtest MAPE/MASE, and an ordering recommendation.

Write a punchy 60-90s product demo video script: hook, problem, demo beats, payoff, CTA — with storyboard, B-roll suggestions, and on-screen copy.

Dependency Agent — runs pip-audit / npm audit / cargo audit after code generation. Produces: a CVE severity report, a license compliance check, a version pinning strategy, and a safe upgrade script. Security-conscious devs' favourite agent.

Detect dependency conflicts, vulnerabilities, and outdated packages across Python and Node. Returns a fix plan and remediation script.

Scan a dependency manifest (requirements.txt, package.json, go.mod, Cargo.toml) for CVEs and severely outdated packages. Returns a severity-ranked report with upgrade commands.

Implement DigiLocker OAuth, document pull, Aadhaar OTP, PAN verify, and eKYC flows (India).

Design RTO/RPO-aware DR plans, failover runbooks, and DR test scripts for cloud systems.

Predict disk usage trends, identify cache directories safe to prune, and warn before 'no space left on device' events.

Auto-detect the best TIMPS agent for any plain-English request and run it immediately. Use this when unsure which agent to pick. Examples: 'my wifi keeps dropping', 'why is my laptop slow', 'broken python environment', 'organize my downloads'.

Compose a production-grade multi-service docker-compose (or compose.yaml) for a project: env, healthchecks, volumes, networks, profiles.

Read raw source code and insert production-quality docstrings in your preferred format (Google, Sphinx, NumPy, JSDoc, Doxygen). Logic is never changed — only documentation is added.

Detect errors and performance issues in shell config files (.zshrc, .bashrc, .gitconfig). Finds slow evals, duplicate PATH entries, syntax errors.

Audit a system or data flow against India's Digital Personal Data Protection Act 2023: consent, notice, data-principal rights, breach, penalties, and a remediation roadmap.

Generate Cloudflare Workers/Vercel Edge/Fastly code with KV storage and geo-routing.

Create Arduino/ESP32/RPi firmware, HAL headers, CMakeLists, and OTA update scripts.

Diagnose broken development environments. Checks Python, Node.js, Docker, Git, PATH integrity, shell config conflicts. Returns exact terminal commands to fix each issue.

Design feature flags with LaunchDarkly/Flagsmith/GrowthBook targeting rules and rollout plans.

Design Flower/FedAvg/PySyft FL pipelines with differential privacy and aggregation strategies.

Scan Downloads and Desktop for clutter: duplicates, large files, unnamed junk. Generates a folder organisation plan and a move script (dry-run — no files moved until user approves).

Generate complete LoRA/MLX/Axolotl/Unsloth fine-tuning pipelines with push-to-hub scripts.

Analyse cloud costs with Infracost, generate rightsizing recommendations and budget alerts.

Identify flaky tests by scanning for timing dependencies, random data, and external HTTP calls. Reads pytest failure cache for history.

Find non-deterministic tests from CI run history. Returns ranked list of flaky tests, suspected causes, and fix suggestions.

Assess an F&B product / packaging against India's FSSAI regulations: license type, label checklist, additive and claim restrictions, and a submission-ready checklist.

Run a comprehensive computer health scan across 6 agents: system optimizer, battery analyst, security guard, network medic, backup sentinel, and update manager. Returns a combined health score and prioritised action list.

Plan and run a chaos-engineering game-day: hypothesis, blast radius, rollback, comms tree, inject schedule, success criteria, and a blameless retrospective template.

Generate an opinionated git workflow: pre-commit hooks, commitlint, lefthook, semantic-release / release-please config, PR template, branch-protection policy, CODEOWNERS.

Write GraphQL SDL schemas, resolvers, DataLoader, subscriptions, and federation config.

Generate GSTR-1/3B filings, e-invoice XML, ITC reconciliation reports (India GST).

Extract translatable strings from source code, generate translation JSON files for multiple locales, and flag RTL/cultural issues.

Detect drift between Terraform / Pulumi / CloudFormation state and live cloud — produces a ranked diff, blast-radius estimate, and a safe terraform plan review.

Triage a batch of emails: flag urgent ones, draft routine replies, identify newsletters to unsubscribe from. Pass a list of email objects with 'from', 'subject', 'body'.

Multi-source log correlation for production incidents. Gathers Docker + app logs, builds a timeline, and generates a triage report.

Turn an active incident (SIEM alert, customer report, CVE disclosure) into a play-by-play response with comms, containment, and a postmortem stub.

India GTM plan: INR pricing tiers, Razorpay integration, GST invoice generation.

Triage a bug report or GitHub issue: detect duplicates, assign severity (critical/high/medium/low), suggest labels and owner. Pass the raw issue title and body text.

Check the status of a previously delegated kernel run by its run_id.

Diagnose Kubernetes workloads: pods, services, ingress, RBAC, resources, network policies. Outputs fix YAML and Helm/Kustomize patches.

Analyse past successful runs and distil prompt improvements for all agents.

Scan source + manifests for OSS license usage (GPL/AGPL/LGPL/SSPL/Commercial), produce a per-dependency report, and a copyleft-compatibility matrix for your distribution model.

List all TIMPS Swarm agents with their purpose, so you can choose the right one for the user's request.

Show all configured LLM providers and which one is currently active. Useful for debugging which model is handling agent calls.

Generate k6/Artillery/Locust load test scripts with ramp-up plans and Grafana dashboards.

Build a fully-local RAG pipeline (Ollama embeddings + Chroma/FAISS + reranker + chat) from a corpus description. Emits ingest, retrieve, generate, evaluate, Dockerfile, Makefile.

Cluster production log errors, identify root cause, and suggest a fix. Handles stack traces, JSON logs, syslog, and journald output.

Read crash logs and system logs, extract stack traces, and explain each crash in plain English. Classifies as app bug / OS bug / hardware / user error. Pass a log file path to analyse a specific log.

Cluster production log errors, identify root cause, suggest a fix. Handles stack traces, JSON logs, syslog, and journald output.

Scaffold a complete Model Context Protocol (MCP) server: tools, resources, prompts, server.py, pyproject, smoke-test client, README.

Scan Photos, Downloads, Desktop for media chaos: unnamed photos, screenshot backlog, oversized videos. Suggests a date-based rename plan and generates ffmpeg compression commands (dry-run).

Extract action items, decisions, and open questions from a meeting transcript. Returns a structured summary with owners and deadlines.

Store, recall, and manage TIMPS run history and contextual memory.

Predict merge conflicts between two branches before merging, with file-level and line-level overlap analysis.

Analyse a codebase for breaking-change APIs given a target library/framework upgrade. Returns a step-by-step migration checklist.

Scaffold React Native/Flutter/Expo apps with push notifications and deep links.

Create evaluation harnesses, adversarial test inputs, and RAGAS config for LLM models.

Monitor an ML model in production: drift, accuracy decay, latency, and cost — with alerting thresholds and a re-training trigger plan.

Configure Stripe/LemonSqueezy products, checkout flows, webhooks, and dunning emails.

Generate Prometheus metric definitions, Grafana dashboard JSON, alerting rules, and instrumentation code snippets.

n8n Workflow Agent — converts a plain-English task description into a complete, importable n8n workflow JSON. Knows all n8n core nodes, integrations (Slack, Gmail, GitHub, Notion, Airtable, HTTP Request, Webhook, Cron, etc.), and expression syntax. Returns the workflow JSON + a visual description of every node.

Diagnose network problems: ping internet/Cloudflare, DNS resolution, WiFi signal, traceroute, open ports. Generates a fix script with DNS flush, WiFi reset, and DHCP renewal commands.

Analyse your observability bill (Datadog / New Relic / Grafana Cloud / Splunk) and recommend cardinality pruning, sampling, log-volume cuts, and tier migrations with projected savings.

Generate a customised onboarding guide: most-touched files, architecture overview, key modules, recent PR walkthroughs.

Detect duplicate code blocks, god classes, magic numbers, and anti-patterns in a codebase.

Generate a realistic, organisation-tailored phishing simulation: emails, landing pages, payload tracking, and a click-through metrics plan. Strictly for authorised training use.

Read a failing CI/CD run and produce a minimal diff to make it green: cache tweaks, image bumps, secret injection, retry policies.

Turn a podcast transcript / outline into SEO-friendly show notes: title, summary, chapter timestamps, key quotes, links, guest bio, and 3 tweet drafts.

Write blameless postmortems: 5 Whys, root causes, P0-P2 action items, runbooks.

Privacy audit: count cookies per browser (Chrome, Firefox, Edge, Brave), list macOS app permissions (camera, mic, location, contacts). Returns a cleanup manifest — no data is deleted until user reviews.

Rewrite prompts with Chain-of-Thought, XML tags, and few-shot examples for better LLM output.

Scan prompts, system messages, and tool schemas for prompt-injection / jailbreak / data-exfiltration risk. Produces a severity-rated report and hardened prompt rewrites.

Review a pull request diff for blockers, suggestions, and nitpicks. Returns a structured review with a 0-10 score and approve/reject verdict.

Identify quantum-vulnerable algorithms, generate Qiskit circuits, and produce NIST PQC migration plans.

Design chunking/embedding/retrieval/reranking RAG pipelines for LlamaIndex or LangChain.

Run a RAGAS-style evaluation of a RAG pipeline: context precision/recall, faithfulness, answer relevance, MRR/NDCG, with a runnable eval harness and an improvement plan.

Build WebSocket/SSE/Redis Pub-Sub/Kafka server+client real-time systems.

Plan and execute an OWASP-style red-team engagement: threat model, adversarial test cases, payloads, runnable exploit harness, and a remediation checklist.

Detect code smells and produce a fully refactored version with a diff summary, complexity scores, and improvement list.

Produce a release plan: SemVer bump, release notes, rollout/canary strategy, rollback gate, customer-comms draft, and stakeholder checklist.

Research Agent — runs BEFORE any coding task. Searches documentation, GitHub issues, Stack Overflow, and the web to gather context so all downstream agents have grounded information. Returns a structured brief: findings, gotchas, recommended approach, sources.

Deep-dive research on any topic. Returns a structured brief: key findings, consensus view, controversies, recommended actions, and sources. Depth options: quick, medium (default), deep.

Generate ROS2 nodes, URDF descriptions, motion planning config, and ISO 10218-1 safety analysis.

Run the full 10-agent SDLC pipeline (product manager → architect → code generator → reviewer → QA → security → performance → devops → docs). Use for software development tasks: writing code, fixing bugs, building APIs, etc.

Generate a CycloneDX / SPDX Software Bill of Materials from a project tree + manifests, integrate syft/trivy if available, and flag license and supply-chain risks.

Scan for hardcoded secrets, generate Vault/AWS Secrets Manager migration code and pre-commit hooks.

Security scan: open network ports, running process anomalies, app camera/microphone/location permissions (macOS TCC). Returns CVSS-rated findings and mitigation commands.

Convert a security finding (CVE, SAST, secret, misconfig) into a step-by-step patch plan with a working code diff and rollback.

Score any agent's output on a 1–10 quality scale, identify weaknesses, and automatically re-run the originating agent with an improved prompt until the score meets the threshold. This closes the quality feedback loop across the entire swarm.

Run Lighthouse CI, generate schema.org JSON-LD, fix Core Web Vitals, build sitemaps.

Generate production-grade service-mesh config (Istio / Linkerd / Consul) — mTLS, traffic policies, retries, circuit-breakers, observability, multi-cluster, and a migration plan.

Decompose epics into sprint tickets with estimates, dependencies, and GitHub Issues payload.

Generate a formatted daily standup report from task lists. Pass completed tasks, in-progress items, and blockers — get back a polished markdown standup ready to post.

Analyse a slow SQL query and return: the optimised rewrite, index recommendations with CREATE INDEX SQL, and an estimated speedup. Supports PostgreSQL, MySQL, SQLite, MSSQL.

Generate an original illustrated children's story for a target age, theme, and moral: outline, full prose, page breaks, illustration prompts, and a printable PDF layout spec.

Diagnose why a computer is slow. Scans: top CPU/RAM processes, startup items, thermal throttling, memory pressure. Returns a report with specific processes to kill and startup items to disable. Generates a dry-run cleanup shell script.

Scan codebase for TODO/FIXME density, cyclomatic complexity hotspots, and deprecated patterns. Returns a prioritised debt repayment plan.

Quantify technical debt via radon complexity + git churn + TODO density; output a debt roadmap.

Refresh a project README, write ADRs, generate a 'Getting Started' guide, an FAQ, and a reference page from the codebase.

Review terraform plan output for destructive changes. Requires explicit approval before allowing apply on high-risk plans.

Generate realistic, edge-case-covering seed / fixture data for any JSON schema, SQL DDL, or plain-text description.

Analyse test suite + git bug history to find untested code paths, suggest property-based tests, and propose a mutation-testing config.

Ingest threat feeds (CVE, MITRE ATT&CK, CISA KEV) and produce a threat brief with detection rules and prioritised mitigations.

Show which AI coding tools are installed and which ones are already connected to the TIMPS Swarm MCP server.

Generate an OpenTelemetry tracing bootstrap: service name, resource attributes, auto-instrumentation, sampling, span/metric/log pipeline, and exporter config.

Monitor keywords across Hacker News, Reddit, GitHub, and Dev.to for spikes in interest. Returns trend direction, sentiment, and a digest summary.

Generate a UI component spec, implementation code, Storybook story, design tokens, and WCAG accessibility audit.

Analyse source code and generate a comprehensive test suite: happy paths, edge cases, error conditions. Returns a complete test file ready to run.

Check for pending updates: macOS/Linux OS, Homebrew packages, global npm packages, pip packages. Returns a prioritised update plan (security first) with a safe ordered shell script.

Generate UPI deep links, QR codes, Razorpay/Cashfree webhook handlers (India UPI).

Design Qdrant/Pinecone/Weaviate/pgvector schemas and integration code.

Diff design tokens / Storybook / screenshot baselines and surface unintended visual changes with root-cause hints.

Design a production voice / phone agent: persona, system prompt, SSML, tool calls, barge-in handling, escalation rules, latency budget, and test utterances.

Interpret wearable / fitness data (HRV, sleep, RHR, steps, SpO2) into an actionable, evidence-based weekly coaching plan with safety guardrails.

Audit smart contracts with Slither/Mythril patterns; generate Solidity code with gas optimisation.

Generate Playwright/Scrapy/httpx scrapers with anti-bot config and data extraction logic.

Multi-engine web search (DuckDuckGo / Brave / SerpAPI / Bing) with deduplication, content extraction, and LLM-synthesised answer with [N] citations. Use for live lookups, news, vendor docs, and competitive intel.

Analyse a set of won/lost deals to surface the real reasons — pricing, champion, timing, competitor, fit — and recommend sales-coaching and product fixes.