Skip to content

Platform Tools (Streams E–I)

Streams E through I cover the broader development ecosystem — stack decisions, agent generation, dev lifecycle tooling, git integration, multi-agent orchestration, governance, and operations.

Quick Reference

ToolDescriptionAvailability
suggest_stackRecommend the best tech stack based on project goals and constraintsIncluded
check_versionsCheck installed versions of frameworks and tools; recommend updatesIncluded
generate_skillGenerate a Claude skill from a spec or codebase patternIncluded
skill_searchSearch for installable Claude skills across multiple registriesIncluded
skill_installInstall a skill from a registry into a project's .claude/skills/ directoryIncluded
detect_agentDetect AI agent capabilities and behaviors in the codebaseIncluded
detect_deprecationsDetect deprecated patterns, APIs, or libraries in useIncluded
generate_sub_agentGenerate a specialized sub-agent configuration for a specific taskIncluded
generate_rulesGenerate project rules and conventions from a specIncluded
generate_teammate_promptGenerate structured agent prompts for an Agent TeamIncluded
suggest_toolingSuggest tools and integrations that would improve development workflowsIncluded
discover_mcpsDiscover available MCP servers and their capabilitiesIncluded
suggest_mcpsSuggest MCP servers based on project needsIncluded
suggest_mcp_serverSuggest appropriate MCP servers for the current projectIncluded
generate_testsGenerate comprehensive test cases and test scenarios from a specIncluded
generate_docsGenerate technical documentation automatically from specsIncluded
generate_ciGenerate GitHub Actions CI workflow configuration for a projectIncluded
generate_planu_ciGenerate a Planu-specific GitHub Actions CI workflowIncluded
transform_codeApply deterministic code transformations using the TypeScript Compiler APIIncluded
integrate_pmIntegrate project management tools and workflows with PlanuIncluded
update_registryAdd a custom framework entry to the local registryIncluded
manage_hooksConfigure and manage git hooks for the projectIncluded
start_hooksStart the file watcher and hook engine for a projectIncluded
stop_hooksStop the file watcher and hook engineIncluded
hook_statusGet the current status of the hook engineIncluded
configure_hooksView or update hook engine configuration at runtimeIncluded
manage_gitManage git operations: branches, commits, push, PRsIncluded
create_pr_from_specCreate a GitHub Pull Request from a specIncluded
review_prAnalyze a Pull Request diff for review findingsIncluded
pr_statusList open Pull Requests with review status and CI checksIncluded
create_releaseCreate a GitHub Release with auto-generated release notesIncluded
generate_changelogGenerate a changelog from merged Pull Requests between two git refsIncluded
orchestrateCoordinate and execute complex multi-step workflowsIncluded
orchestrate_runtimeLaunch and coordinate parallel agents with configurable topologiesIncluded
orchestrate_agentsOrchestrate multiple AI agents for parallel or sequential task executionIncluded
agent_swarm_statusGet status of orchestration sessions: agent progress, task status, metrics, and costIncluded
validate_team_resultsValidate the combined results from all teammates after parallel executionIncluded
manage_contextManage and optimize context usage across specs and projectsIncluded
context_budgetView and manage your context token budget for AI-assisted developmentIncluded
token_intelligenceAdvanced token analytics: cost tracking, budget monitoring, estimation accuracy, trend detection, and anomaly alertsIncluded
consult_docsSearch and consult project documentation without leaving the CLIIncluded
migrate_techPlan and execute a technology migrationIncluded
ecosystem_statusScan external ecosystem sources for new capabilities, patterns, and integration opportunitiesIncluded
generate_infrastructureGenerate Infrastructure-as-Code (IaC) from a specIncluded
configure_rolesEnable or disable Role-Based Access Control (RBAC) for a projectIncluded
assign_roleAssign a role to a user for a project or a specific specIncluded
configure_workersConfigure background analysis workers: enable/disable, set cooldown, set priorityIncluded
worker_statusShow the status of background analysis workersIncluded
configure_llm_providersConfigure LLM providers for multi-provider supportIncluded
llm_provider_statusShow status of all configured LLM providers including health and cost trackingIncluded
configure_observabilityConfigure OpenTelemetry (OTel) tracing and metricsIncluded
manage_pluginsList, install, activate, deactivate, uninstall, or get info about Planu pluginsIncluded
scaffold_pluginGenerate a starter plugin project with manifest, entry point, tests, and examplesIncluded
token_usageView token consumption metrics, cache performance, and optimization savingsIncluded
submit_feedbackSubmit a bug report, feature request, or suggestionIncluded
triage_feedbackView and prioritize pending user feedback grouped by similarityIncluded
resolve_feedbackMark feedback as resolved, linking it to the spec and PR that fixed itIncluded
registry_searchSearch the Planu spec registry for reusable specsIncluded
registry_installInstall a spec from the Planu registry into a local projectIncluded
configure_oauthCheck or test OAuth 2.1 Bearer token authenticationIncluded

Stream E — Stack & Agents

suggest_stack

Recommend the best tech stack based on project goals and constraints.

When to use: At project inception when the team is choosing between competing technologies, or when validating a stack choice against project constraints.

Prompt: "Suggest a stack for a real-time collaborative document editor for project proj_abc123"

check_versions

Check installed versions of frameworks and tools; recommend updates.

What it checks:

  • Current vs latest stable versions
  • Known CVEs and security advisories
  • Deprecated packages
  • Breaking change warnings for major version gaps
Prompt: "Check dependency versions for project proj_abc123"

generate_skill

Generate a Claude skill from a spec or codebase pattern.

Supported platforms: Claude Code (CLAUDE.md skills), Cursor (.cursorrules fragments), Windsurf, Copilot instructions, custom agent definitions.

Prompt: "Generate a Claude Code skill for the authentication workflow in project proj_abc123"

Search for installable Claude skills across multiple registries.

When to use: When you want to find community-maintained or Planu-curated skills relevant to your stack or workflow.

Prompt: "Search for TypeScript skills in the Planu skill registry"

skill_install

Install a skill from a registry into a project's .claude/skills/ directory.

When to use: After finding a skill with skill_search, to add it to your project so agents can use it automatically.

Prompt: "Install the typescript-expert skill into project proj_abc123"

detect_agent

Detect AI agent capabilities and behaviors in the codebase.

When to use: When you need to generate agent-specific artifacts (rules files, skills, memory files) and want Planu to auto-select the right format.

Prompt: "Detect what AI agent is configured at /workspace/my-app"

detect_deprecations

Detect deprecated patterns, APIs, or libraries in use.

When to use: Before a major version upgrade, or as part of a quarterly tech debt review. Scans project source files with no network required.

Prompt: "Detect deprecated APIs in project proj_abc123"

generate_sub_agent

Generate a specialized sub-agent configuration for a specific task.

What it produces: A structured agent definition with capabilities, resource ownership, handoff protocol, and communication contract — ready for use with orchestrate_agents.

Prompt: "Generate a sub-agent for the database migration task in spec SPEC-011 for project proj_abc123"

generate_rules

Generate project rules and conventions from a spec — .cursorrules, CLAUDE.md, Copilot instructions, or Windsurf rules — derived from the project's Constitution and detected patterns.

Prompt: "Generate a CLAUDE.md for project proj_abc123"
Prompt: "Generate .cursorrules for project proj_abc123"

generate_teammate_prompt

Generate structured agent prompts for an Agent Team.

When to use: When setting up parallel agent execution with orchestrate_agents or orchestrate_runtime. Each teammate gets a focused, non-overlapping prompt with explicit file ownership.

Prompt: "Generate teammate prompts for specs SPEC-012 through SPEC-015 in project proj_abc123"

Stream F — Dev Lifecycle

suggest_tooling

Suggest tools and integrations that would improve development workflows.

Coverage:

  • CI/CD (GitHub Actions, GitLab CI, CircleCI, etc.)
  • Testing (unit, integration, E2E, visual regression)
  • Monitoring and observability (OpenTelemetry, Datadog, Grafana)
  • AI-native tooling (LLM observability, semantic cache, prompt management, guardrails)
  • Frontend performance tooling (Lighthouse CI, visual regression, component libraries)
Prompt: "Suggest tooling for project proj_abc123"

discover_mcps

Discover available MCP servers and their capabilities.

When to use: When you want to understand which MCP servers are configured in the current environment.

Prompt: "What MCP servers are available in my environment?"

suggest_mcps

Suggest MCP servers based on project needs.

When to use: When starting a new project and wanting recommendations for MCP servers that match your detected stack and workflow patterns.

Prompt: "Suggest MCP servers for project proj_abc123"

suggest_mcp_server

Suggest appropriate MCP servers for the current project.

When to use: When you need a focused recommendation for a specific capability gap (e.g., "I need an MCP server for database operations").

Prompt: "Suggest an MCP server for managing PostgreSQL for project proj_abc123"

generate_tests

Generate comprehensive test cases and test scenarios from a spec's acceptance criteria.

What it produces:

  • Test plan with test case names mapped to acceptance criteria
  • Test file stubs in your project's test framework (vitest, pytest, go test, JUnit, etc.)
  • Coverage targets per criterion
  • Mocking strategy for external dependencies
Prompt: "Generate tests for spec SPEC-003 in project proj_abc123"

generate_docs

Generate technical documentation automatically from specs — API docs, README sections, architecture overviews.

Prompt: "Generate API documentation for spec SPEC-007 in project proj_abc123"
Prompt: "Generate a README section for the authentication module from spec SPEC-001"

generate_ci

Generate GitHub Actions CI workflow configuration for a project.

Supported: GitHub Actions, GitLab CI, CircleCI, Jenkins, Azure Pipelines, Bitbucket Pipelines.

Prompt: "Generate a GitHub Actions CI pipeline for project proj_abc123"

generate_planu_ci

Generate a Planu-specific GitHub Actions CI workflow that validates specs, checks drift, and enforces quality gates as part of your CI pipeline.

Prompt: "Generate a Planu CI check for my GitHub Actions pipeline in project proj_abc123"

transform_code

Apply deterministic code transformations using the TypeScript Compiler API.

When to use: When migrating code patterns (class-based to functional, callbacks to async/await), refactoring to match spec requirements, or applying consistent patterns across a codebase.

Prompt: "Transform the auth module from callbacks to async/await in project proj_abc123"

integrate_pm

Integrate project management tools and workflows with Planu — create issues, epics, or tasks from specs.

Supported: Jira, Linear, GitHub Issues, GitLab Issues, Trello, Notion.

Prompt: "Create a GitHub issue for spec SPEC-004 in project proj_abc123"
Prompt: "Create a Linear task for spec SPEC-009 in project proj_abc123"

update_registry

Add a custom framework entry to the local registry.

When to use: When your project uses a framework not recognized by Planu's built-in catalog — adds it so suggest_stack, check_versions, and suggest_tooling can use it.

Prompt: "Add my custom framework 'acme-framework' v2.1 to the local registry for project proj_abc123"

manage_hooks

Configure and manage git hooks for the project — pre-commit, post-merge, pre-push — linked to spec validation and quality gates.

When to use: When setting up or modifying git hooks that enforce SDD workflow rules.

Prompt: "Set up pre-commit hooks for spec validation in project proj_abc123"

start_hooks

Start the file watcher and hook engine for a project.

When to use: To activate real-time hook processing during an active development session.

Prompt: "Start the hook engine for project proj_abc123"

stop_hooks

Stop the file watcher and hook engine.

Prompt: "Stop the hook engine for project proj_abc123"

hook_status

Get the current status of the hook engine — whether it is running, which hooks are active, and recent trigger history.

Prompt: "Check hook engine status for project proj_abc123"

configure_hooks

View or update hook engine configuration at runtime without restarting.

When to use: To change cooldown intervals, enable/disable specific hooks, or adjust hook priorities without stopping and restarting the engine.

Prompt: "Configure hooks to increase cooldown to 60 seconds for project proj_abc123"

Stream G — Git & CI

manage_git

Manage git operations: branches, commits, push, PRs — linked to specs for full traceability.

Prompt: "Create a branch for spec SPEC-005 in project proj_abc123"
Prompt: "Generate a PR description for spec SPEC-005 in project proj_abc123"
Prompt: "Generate a changelog from specs SPEC-001 through SPEC-010 for project proj_abc123"

create_pr_from_spec

Create a GitHub Pull Request from a spec — auto-fills title, body, labels, and linked spec metadata.

When to use: After implementing a spec and wanting to open a PR with full SDD traceability (spec ID, acceptance criteria, implementation notes).

Prompt: "Create a PR from spec SPEC-012 in project proj_abc123"

review_pr

Analyze a Pull Request diff for review findings — checks for spec compliance, code patterns, and quality issues.

When to use: Before approving a PR, or as an automated step in CI to catch spec drift early.

Prompt: "Review PR #88 for project proj_abc123"

pr_status

List open Pull Requests with review status and CI checks.

When to use: To get a quick overview of all in-flight PRs and their blocking status before a planning or standup session.

Prompt: "Show PR status for project proj_abc123"

create_release

Create a GitHub Release with auto-generated release notes derived from merged specs and PRs.

Prompt: "Create a release v2.1.0 for project proj_abc123"

generate_changelog

Generate a changelog from merged Pull Requests between two git refs.

When to use: Before a release to compile a human-readable summary of all changes since the last tag.

Prompt: "Generate a changelog from v2.0.0 to HEAD for project proj_abc123"

Stream H — Orchestration & Context

orchestrate

Coordinate and execute complex multi-step workflows — register agents, lock resources, distribute tasks, and manage parallel execution state.

Prompt: "Register an agent for spec SPEC-011 and lock the auth module in project proj_abc123"

orchestrate_runtime

Launch and coordinate parallel agents with configurable topologies (fan-out, pipeline, consensus).

When to use: When running 3+ independent implementation tasks in parallel and needing centralized state management, conflict detection, and result aggregation.

Prompt: "Launch a fan-out orchestration for specs SPEC-020 through SPEC-025 in project proj_abc123"

orchestrate_agents

Orchestrate multiple AI agents for parallel or sequential task execution — declare capabilities, route tasks, generate handoff contexts, and detect file-ownership conflicts before execution.

Prompt: "Orchestrate agents for specs SPEC-012, SPEC-013, SPEC-014 in project proj_abc123"

agent_swarm_status

Get status of orchestration sessions: agent progress, task status, metrics, and cost.

When to use: During or after a parallel agent session to monitor which agents are blocked, completed, or failing.

Prompt: "Show swarm status for orchestration session orch_abc123"

validate_team_results

Validate the combined results from all teammates after parallel execution — checks for conflicts, missing outputs, and spec compliance.

When to use: After orchestrate_agents completes, before merging parallel branches.

Prompt: "Validate team results for orchestration session orch_abc123 in project proj_abc123"

manage_context

Manage and optimize context usage across specs and projects — save, load, and curate persistent context notes.

When to use: When you need to preserve important architectural decisions, external constraints, or team agreements that aren't captured in specs.

Prompt: "Save context: we decided to use Postgres 16 with pgvector for all AI features in project proj_abc123"
Prompt: "What context have we saved for project proj_abc123?"

context_budget

View and manage your context token budget for AI-assisted development.

Actions:

  • status — Check current token usage, alert level, and top consumers
  • configure — Set context window size and warning/critical thresholds
  • checkpoint — Save session state (decisions, files, specs, git status) for continuity
  • restore — Load a previous checkpoint to resume work in a new session
  • list-checkpoints — Show available session checkpoints
  • optimize — Get delegation suggestions to reduce context pressure
Prompt: "Check my context budget status"
Prompt: "Save a checkpoint of this session"
Prompt: "What should I delegate to subagents?"

token_intelligence

Advanced token analytics: cost tracking, budget monitoring, estimation accuracy, trend detection, and anomaly alerts.

When to use: When running long or expensive agent sessions and needing visibility into where tokens are being spent, whether estimates are accurate, and where to cut costs.

Prompt: "Show token intelligence report for project proj_abc123"
Prompt: "Alert me if any session exceeds 50k tokens"

consult_docs

Search and consult project documentation without leaving the CLI — fetches and summarizes relevant docs in context.

Prompt: "Consult React docs about concurrent rendering for project proj_abc123"
Prompt: "What does the FastAPI docs say about dependency injection?"

migrate_tech

Plan and execute a technology migration — analyze source codebase, map equivalences, generate migration plan, validate functional parity, and plan database migration.

When to use: When moving from one framework to another (Express → Fastify, Django → FastAPI, Create React App → Vite), upgrading a major dependency, or migrating databases.

Prompt: "Plan a migration from Express to Fastify for project proj_abc123"
Prompt: "Migrate from class-based to functional React components in project proj_abc123"

ecosystem_status

Scan external ecosystem sources for new capabilities, patterns, and integration opportunities relevant to the project's stack.

When to use: At the start of a new phase or quarterly, to stay current with framework releases, new MCP servers, and emerging tooling patterns.

Prompt: "Run ecosystem status scan for project proj_abc123"

generate_infrastructure

Generate Infrastructure-as-Code (IaC) from a spec — Terraform, Pulumi, CDK, Helm charts, or Docker Compose.

When to use: When a spec includes infrastructure requirements (databases, queues, caches, serverless functions) and you want reproducible IaC instead of manual provisioning.

Prompt: "Generate Terraform for the database spec SPEC-030 in project proj_abc123"
Prompt: "Generate a Helm chart for the API deployment in spec SPEC-031"

Stream I — Governance & Operations

configure_roles

Enable or disable Role-Based Access Control (RBAC) for a project — define which roles can create, approve, or implement specs.

Prompt: "Enable RBAC for project proj_abc123 with roles: architect, developer, reviewer"

assign_role

Assign a role to a user for a project or a specific spec.

When to use: When onboarding a new team member and defining their permissions within the SDD workflow.

Prompt: "Assign the 'developer' role to alice@example.com in project proj_abc123"

configure_workers

Configure background analysis workers: enable/disable, set cooldown, set priority.

Workers available: drift detector, convention scanner, spec health monitor, dependency auditor, auto-memory injector.

Prompt: "Enable the drift detector worker with a 30-minute cooldown for project proj_abc123"
Prompt: "Set the convention scanner to high priority for project proj_abc123"

worker_status

Show the status of background analysis workers — which are running, last run time, and recent findings.

Prompt: "Show worker status for project proj_abc123"

configure_llm_providers

Configure LLM providers for multi-provider support — add API keys, set routing rules, and define fallback chains.

Supported providers: Anthropic, OpenAI, Google Gemini, Mistral, Groq, Ollama (local).

Prompt: "Add OpenAI as a secondary LLM provider for project proj_abc123"

llm_provider_status

Show status of all configured LLM providers including health, latency, and cost tracking.

Prompt: "Show LLM provider status for project proj_abc123"

configure_observability

Configure OpenTelemetry (OTel) tracing and metrics — export spans to Jaeger, Grafana Tempo, Datadog, or Honeycomb.

When to use: When running Planu in a team environment and wanting distributed tracing for agent sessions and tool calls.

Prompt: "Configure OpenTelemetry export to Grafana Tempo for project proj_abc123"

manage_plugins

List, install, activate, deactivate, uninstall, or get info about Planu plugins.

When to use: When extending Planu with custom tools or third-party integrations packaged as plugins.

Prompt: "List all installed plugins for project proj_abc123"
Prompt: "Install the linear-sync plugin"
Prompt: "Deactivate the jira-integration plugin"

scaffold_plugin

Generate a starter plugin project with manifest, entry point, tests, and examples.

When to use: When building a custom Planu plugin and wanting a fully structured starting point that follows the plugin API contract.

Prompt: "Scaffold a new Planu plugin called 'notion-sync' for project proj_abc123"

token_usage

View token consumption metrics, cache performance, and optimization savings.

When to use: To understand how many tokens your team is spending, which tools are most expensive, and how much the cache is saving.

Prompt: "Show token usage for project proj_abc123"

submit_feedback

Submit a bug report, feature request, or suggestion about Planu directly from your conversation.

Inputs:

  • typebug, feature, suggestion, or dx
  • title — Short summary of the feedback
  • description — Detailed explanation
  • context (optional) — Relevant project ID, spec ID, or tool name

Feedback is stored locally and sent anonymously to Planu's central tracker.

Prompt: "I want to report a bug with the estimate tool"
Prompt: "Submit a feature request: I'd like create_spec to support importing from Notion"
Prompt: "I have a suggestion about the onboarding flow"

triage_feedback

View and prioritize pending user feedback grouped by similarity — helps teams decide what to act on next.

What it produces:

  • Grouped feedback items (similar bugs and requests merged using similarity scoring)
  • Priority score per group (frequency × severity)
  • Suggested action per group: create_spec | merge | dismiss
Prompt: "Triage all pending feedback"
Prompt: "Show me pending bug reports"
Prompt: "Triage feature requests for project proj_abc123"

resolve_feedback

Mark feedback as resolved, linking it to the spec and pull request that addressed it — creating full traceability from user feedback to shipped fix.

Inputs:

  • feedbackId — ID of the feedback item to resolve
  • specId (optional) — The spec that addressed this feedback
  • prUrl (optional) — URL of the pull request that shipped the fix
  • notes (optional) — Resolution notes or context
Prompt: "Mark feedback fb_001 as resolved, linked to SPEC-042"
Prompt: "Resolve feedback fb_003 with PR https://github.com/org/repo/pull/88"

Search the Planu spec registry for reusable specs — find community-contributed specs that match your feature needs.

Prompt: "Search the Planu registry for authentication specs"
Prompt: "Find specs related to payment processing in the registry"

registry_install

Install a spec from the Planu registry into a local project.

When to use: After finding a relevant spec with registry_search, to add it as a starting point for your own implementation.

Prompt: "Install spec planu/stripe-checkout into project proj_abc123"

configure_oauth

Check or test OAuth 2.1 Bearer token authentication for the Planu HTTP transport.

When to use: When running Planu as an HTTP server (not stdio) and you need to protect it with an API key or HMAC-signed JWT tokens.

What it produces:

  • Saves OAuth configuration (mode: api-key or jwt)
  • Returns the environment variable names to set (SDD_OAUTH_API_KEY or SDD_OAUTH_JWT_SECRET)
  • Validation summary confirming the transport is now protected
Prompt: "Configure OAuth for my Planu HTTP server using API key mode"

See Also

Join the communityAsk questions, share feedback, and connect with other developers using Planu.
Join Discord