jaan.to Roadmap

See vision.md for philosophy and concepts

Overview

Phase	Focus	Status
1-3	Foundation, migration, customization	Done (see CHANGELOG.md)
4	Development workflow	Planned
5	Detect & Knowledge Pack (6 skills)	Done
6	Role skills (58 shipped across 17 role prefixes)	In Progress
7	MCP connectors (1 full + 3 docs-only / 24 cataloged)	In Progress
8	Testing and polish	Planned
9	Distribution & CLI transformation (1/7 done)	Partial

Version History

For complete release history, see CHANGELOG.md.

Latest: v7.7.1 (58 skills)

Unreleased

Role Orchestrator Skills — 6 per-role orchestrator skills (/pm, /ux, /dev, /qa, /devops, /sec) using Claude Code Agent Teams. Each orchestrator coordinates all sub-skills within its role via dynamic discovery (sub-skills.md). Update team-ship to delegate to orchestrators as meta-orchestrator with backward-compatible fallback. → details
Skill Lifecycle Automation — 5 workflow automation skills discovered via pm-skill-discover (est. ~333 min/week savings):
- dev-adapter-sync (Must, Quick Win) — Mirror skill files from skills/ to adapters/codex/ automatically. High priority: manual build-codex-skillpack.sh is a repeated friction point in every PR
- dev-skill-batch-update (Must) — Apply uniform changes across all SKILL.md files at once
- dev-docs-sync (Should) — Auto-sync CHANGELOG, roadmap, DEPENDENCIES, READMEs, marketplace.json
- qa-skill-validate (Should) — Validate skills, diagnose failures, auto-fix
- devops-adapter-rebuild (Could) — Rebuild codex adapter after changes

v7.7.1 — 2026-03-15

QA Performance Optimization — Apply research #86/#87 findings across 7 qa-* skills and 3 reference docs (parallelization, incremental mutation, coverage tool selection, fail-fast, batch generation, early-termination)
QA skill description compliance — Fix 3 descriptions exceeding 120-char limit, remove 4 uncited percentage claims

v7.7.0 — 2026-03-05

pm-sprint-plan skill — Assess project progress and build prioritized sprint plan from roadmap gaps with progress matrix and execution queue
team-sprint skill — Full sprint cycle orchestrator from planning to PR, chaining pm-sprint-plan and team-ship --track sprint
team-ship --track sprint — New sprint track that reads sprint plan artifacts and dynamically spawns Agent Teams based on execution queue
58 skills total (was 56 in v7.6.1)

v7.6.0 — 2026-02-26

v7.5.1 — 2026-02-25

Security hardening (6 rounds) — Closed multiple prompt injection bypass vectors in pre-tool security gate
Layered prompt injection defenses across all distributed skills
pm-changelog-rewrite skill — New changelog rewrite and maintenance workflows
Enhanced pm-roadmap-update and release-iterate-changelog with file discovery, auto-commit, issue refs
53 skills total

v7.5.0 — 2026-02-24

Windows NTFS compatibility fix (#157) — Replace jaan-to: colon prefix with jaan-to- dash prefix in all learn/template filenames, auto-migration in bootstrap.sh with three-branch collision handling
TDD/BDD/AI orchestration quality skills — 4 new skills + 15 enhanced, from research doc 76 (qa-test-mutate, qa-tdd-orchestrate, qa-contract-validate, qa-quality-gate + 11 enhanced skills)
qa-issue-validate skill — Validate GitHub/GitLab issues against codebase with RCA and threat scanning
qa-issue-report skill — Report issues to any GitHub/GitLab repository with smart context analysis
Convert internal roadmap skills to generic pm-* skills — pm-roadmap-add and pm-roadmap-update replace internal roadmap-add/roadmap-update
pm-skill-discover skill — Detect repeated workflow patterns and suggest skills to automate them
53 skills total (was 45 in v7.1.0)

v7.4.0 — 2026-02-22

Context7 MCP integration + /dev-docs-fetch skill — library docs fetch, smart caching, auto-detect from tech.md → details
Dual-runtime MCP support — Codex config.toml auto-configuration, validate-mcp-servers.sh parity checks, CI dual-runtime MCP gate
MCP infrastructure hardening — config update safety, .mcp.json build distribution, stale count fixes

v7.3.0 — 2026-02-21

Multi-runtime Codex support — dual-runtime governance, single-source build targets, codex runner, installer-first skillpack, CI dual-runtime enforcement
Codex runner fixes — zero-argument handling, .agents/ native discovery path
Website hero install UX redesign — clipboard buttons, local logos, improved layout

v7.2.0 — 2026-02-19

Security audit remediation — 13 findings fixed, set -euo pipefail enforced, skill permissions narrowed
Automated security enforcement — scripts/validate-security.sh, CI gate, /jaan-release + /jaan-issue-review integration
jaan-init Co-Authored-By attribution for git commits and PRs (#109)
pm-prd-write output readability, document flow, RTL support (#141)
Agent Skills open standard compatibility — All 44 skills compliant with agentskills.io spec (license, compatibility, trigger phrases, marketplace discovery, naming spec, E2E tests, CI enforcement)
team-ship agent teams orchestration skill — Role-based AI teammates for spec-to-ship pipeline
Compliance hardening — Skill Alignment sections (44 skills), Definition of Done (5 skills), hardcoded path sanitization

Phase 4: Development Workflow

Details: tasks/development-workflow.md | tasks/lsp-support.md

Project constitution document (context/constitution.md) — 9 immutable development principles
~~Complexity tracking in outputs~~ — Deferred: overlaps with existing LEARN.md feedback system
LSP support — Bundle TypeScript + Python language server configs, make skills LSP-aware

Phase 5: Detect & Knowledge Pack (6 skills) — DONE

Details: tasks/role-skills/detect.md | Completed in v3.23.0–v6.1.0 — see CHANGELOG.md

6 detect skills shipped: detect-dev, detect-design, detect-writing, detect-product, detect-ux, detect-pack. Pipeline produces $JAAN_OUTPUTS_DIR/detect/ knowledge with SARIF-like evidence and confidence scoring.

Phase 6: Role Skills (142 across 17 role prefixes)

Details: tasks/role-skills.md

142 skills cataloged across 17 role prefixes. Quick-win skills (no MCP required) are built first, followed by advanced skills that depend on MCP connectors.

All new skills must follow v3.0.0 patterns: $JAAN_* environment variables, template variables, tech stack integration, and pass /skill-update validation.

Role	Total	Quick Wins	Advanced	File
PM	24	18	6	pm.md
DEV	19	15	4	dev.md
QA	12	8	4	qa.md
DATA	14	8	6	data.md
GROWTH	15	9	6	growth.md
UX	21	16	5	ux.md
SEC	4	4	0	sec.md
DELIVERY	8	8	0	delivery.md
SRE	9	5	4	sre.md
SUPPORT	8	8	0	support.md
RELEASE	8	8	0	release.md

Priority (by research rank): qa-test-cases (#1), data-sql-query (#2), ux-research-synthesize (#8), qa-bug-report (#10), growth-meta-write (#12). Full priority list in tasks/role-skills.md.

Completed work: 56 skills shipped across v4.0.0–v7.6.0 — see CHANGELOG.md for details.

Phase 7: MCP Connectors (24 documented)

Details: tasks/mcp-connectors.md

MCP connectors provide real system context to skills. Skills stay generic; MCP provides per-product data from actual tools.

Core MCPs (11): GA4, GitLab, Jira, Figma, GSC, Clarity, Sentry, BigQuery, Playwright, OpenAPI, dbt Cloud
Extended MCPs (13): Notion, Slack, GitHub, Linear, Mixpanel, Confluence, Snowflake, PostgreSQL, Ahrefs, Semrush, LambdaTest, Google Drive, Memory
Infrastructure: Deferred loading for token savings, Context7 integration, model routing per skill type
Context7 MCP integration + /dev-docs-fetch skill — library docs fetch, smart caching, auto-detect from tech.md → details

Phase 8: Testing and Polish

E2E test framework in tests/ with mocked MCP responses
JSON export alongside markdown for all skill outputs
External notifications (Slack integration)
[BLOCKER] Fix learn-report skill hook script for macOS compatibility (Bash 3.2 — needs 4+ for declare -A)

Phase 9: Distribution & CLI Transformation

Details: tasks/distribution.md | tasks/cli-transformation.md

Standalone CLI (`jaan-to` npm package)

Build jaan-to as an independent CLI app using the Claude Agent SDK (TypeScript). Dual distribution: plugin stays for Claude Code users, CLI reaches everyone else. Skills, templates, learning, and context files are shared between both runtimes.

Phase A: MVP CLI — SKILL.md parser, config loader, env var resolver, jaan-to init, jaan-to run <skill> → details
Phase B: Feature parity — Learning merge, template variables, subagents, hooks, shell script ports
Phase C: CLI-native features — CI/CD mode (--ci), batch mode, progress UI, session management
Phase D: Multi-model — Provider abstraction layer, OpenAI/Gemini/Ollama support

Cross-Agent Compatibility

Agent Skills open standard compatibility (agentskills.io) — marketplace.json discovery, all 44 skills compliant
Multi-agent compatibility research (Cursor, Copilot, Windsurf, Gemini)

Public Presence

Public documentation site and branding guidelines
npm package publication (jaan-to)

Quick Reference

Commands

Command	Description
`/pm-prd-write`	Generate PRD from initiative
`/pm-research-about`	Deep research or add file/URL to index
`/pm-story-write`	Generate user stories with Given/When/Then AC
`/data-gtm-datalayer`	Generate GTM tracking code
`/detect-dev`	Engineering audit with scored findings
`/detect-design`	Design system detection with drift findings
`/detect-writing`	Writing system extraction with tone scoring
`/detect-product`	Product reality extraction with evidence
`/detect-ux`	UX audit with journey/pain-point findings
`/detect-pack`	Consolidate detect outputs into knowledge index
`/frontend-task-breakdown`	Frontend task breakdown with component hierarchy
`/backend-task-breakdown`	Backend task breakdown from PRDs
`/frontend-design`	Frontend component design
`/backend-data-model`	Generate data model docs with constraints, indexes, and migrations
`/backend-api-contract`	Generate OpenAPI 3.1 contracts from API entities
`/backend-pr-review`	Review backend PRs for security, performance, and code quality
`/backend-scaffold`	Generate production-ready backend code from specs
`/frontend-scaffold`	Convert designs to React/Next.js scaffold code
`/ux-heatmap-analyze`	Analyze heatmap CSV + screenshots
`/ux-microcopy-write`	Multi-language UX microcopy
`/ux-research-synthesize`	Synthesize UX research findings
`/ux-flowchart-generate`	Generate Mermaid flowcharts from PRD/docs/codebase
`/qa-test-cases`	Generate test cases from PRDs
`/wp-pr-review`	Review WordPress plugin PRs for security and standards
`/dev-project-assemble`	Wire scaffold outputs into runnable project structure
`/backend-service-implement`	Generate service implementations from scaffold stubs
`/qa-test-generate`	Generate runnable Vitest/Playwright tests from BDD cases
`/sec-audit-remediate`	Generate security fixes from detect-dev findings
`/dev-output-integrate`	Copy generated outputs into project with config merging
`/devops-infra-scaffold`	Generate CI/CD, Dockerfiles, deployment configs
`/devops-deploy-activate`	Activate deployment pipeline with secrets and platform provisioning
`/qa-test-run`	Execute tests, diagnose failures, auto-fix, generate coverage reports
`/qa-test-mutate`	Run mutation testing and generate survivor reports
`/qa-tdd-orchestrate`	Orchestrate RED/GREEN/REFACTOR TDD cycle with isolated agents
`/qa-contract-validate`	Validate API contracts with Spectral, oasdiff, Prism, Schemathesis
`/qa-quality-gate`	Compute composite quality score from upstream skill outputs
`/qa-issue-validate`	Validate GitHub/GitLab issues against codebase with RCA
`/qa-issue-report`	Report issues to any GitHub/GitLab repository
`/pm-skill-discover`	Detect repeated workflow patterns and suggest skills
`/frontend-story-generate`	Generate CSF3 Storybook stories with CVA variant detection
`/frontend-visual-verify`	Visually verify components using Playwright MCP
`/frontend-component-fix`	Diagnose and fix UI bugs with output-only safety model
`/dev-verify`	Validate build pipeline and running services with health checks
`/dev-docs-fetch`	Fetch and cache library docs via Context7 MCP (Phase 7)
`/release-iterate-changelog`	Generate changelog with user impact and support guidance
`/jaan-init`	Initialize jaan-to for a project
`/jaan-issue-report`	Report issues to jaan-to GitHub repo
`/skill-create`	Create new skill with wizard
`/skill-update`	Update existing skill
`/docs-create`	Create documentation with templates
`/docs-update`	Audit and update stale documentation
`/learn-add`	Add lesson to skill's LEARN.md
`/pm-roadmap-add`	Add prioritized items to project roadmap
`/pm-roadmap-update`	Review and maintain project roadmap
`/learn-report`	View learning system dashboard

Key Paths

Path	Purpose
`skills/`	Skill definitions (plugin-relative)
`jaan-to/config/settings.yaml`	Project configuration (v3.0.0+)
`jaan-to/context/`	Context templates (project-relative)
`jaan-to/templates/`	Output templates (project-relative)
`jaan-to/outputs/`	Generated outputs (project-relative)
`jaan-to/learn/`	Learning files (project-relative)
`.claude-plugin/plugin.json`	Plugin manifest
`docs/roadmap/tasks/`	Task details

Skill Naming

Pattern	Example
Role-based: `{role}-{domain}-{action}`	`/pm-prd-write`
Internal: `{domain}-{action}`	`/skill-create`
Directory: `skills/{skill-name}/`	`skills/pm-prd-write/`

Overview​

Version History​

Unreleased​

v7.7.1 — 2026-03-15​

v7.7.0 — 2026-03-05​

v7.6.0 — 2026-02-26​

v7.5.1 — 2026-02-25​

v7.5.0 — 2026-02-24​

v7.4.0 — 2026-02-22​

v7.3.0 — 2026-02-21​

v7.2.0 — 2026-02-19​

Phase 4: Development Workflow​

Phase 5: Detect & Knowledge Pack (6 skills) — DONE​

Phase 6: Role Skills (142 across 17 role prefixes)​

Phase 7: MCP Connectors (24 documented)​

Phase 8: Testing and Polish​

Phase 9: Distribution & CLI Transformation​

Standalone CLI (jaan-to npm package)​

Cross-Agent Compatibility​

Public Presence​

Quick Reference​

Commands​

Key Paths​

Skill Naming​