org-plan

Plugin gestalt in dyne/gestalt-agents. Commands read from the repo's manifest — today.

Skills (20)

• coding

  Write, review, or refactor application code and architecture. Use for features, APIs, DDD, bounded contexts, aggregates, Clean/Hexagonal Architecture, vertical slices, ports/adapters, repositories, CQRS, events, and monolith decomposition. Follow existing conventions; apply deeper patterns only when

• context-mode

  Route large or uncertain command, file, test, log, API, documentation, browser, and MCP output through context-mode so only derived findings enter the conversation. Use for output analysis, filtering, counting, searching, indexing, or summarizing; keep edits and governing instruction reads on native

• ctx-doctor

  Diagnose context-mode runtime, dependency, spawned MCP handshake, hook, FTS5, plugin registration, version, and startup problems. Trigger when the user invokes ctx-doctor or reports that context-mode is missing, unhealthy, or failing to start.

• ctx-index

  Index a local file or directory into context-mode's FTS5 knowledge base for focused later retrieval. Trigger when the user invokes ctx-index or wants persistent search over project files without rereading raw content.

• ctx-insight

  Open the hosted context-mode Insight engineering analytics dashboard. Trigger when the user invokes ctx-insight or asks to open Insight, view productive-rate analytics, retry waste, blockers, or team role views.

• ctx-purge

  Permanently delete context-mode data for one session or an entire project. Trigger only when the user invokes ctx-purge or explicitly asks to wipe indexed content, session records, events, or context-mode statistics.

• ctx-search

  Search context-mode's FTS5 knowledge base for indexed project files, documentation, or session memory. Trigger when the user invokes ctx-search or asks focused follow-up questions about previously indexed content.

• ctx-stats

  Show context-mode token use, context savings, and per-tool activity for the current session. Trigger when the user invokes ctx-stats or asks how much context-mode saved. This operation is read-only.

• ctx-upgrade

  Repair the locally shipped context-mode installation, runtime bundles, hooks, settings, and plugin registration. Trigger when the user invokes ctx-upgrade or asks to repair context-mode. It never updates from upstream.

• development-testing

  Implement a defined behavior, create or update coherent automated tests for its observable contract, and bring the relevant suite to green. Use when implementation and tests must agree before completion; a red-first test is not required unless another governing workflow requires it.

• org-plan

  Create or execute durable Org plans when explicitly requested or when implementation work needs cross-session persistence, L1/L2 milestones, assigned skills or owners, supervised review and evidence gates, per-milestone commits, or mobile attention. Do not use for ordinary bounded single-session tas

• playwright

  Use Gestalt-managed Playwright MCP for website interaction, inspection, screenshots, and browser verification when the host connection is ready.

• scripting

  Write, review, debug, or refactor Bash/POSIX shell scripts, commands, CI/CD shell steps, and Linux/macOS automation. Covers quoting, pipelines, options, exit codes, traps, temporary files, processes, portability, and idempotency. Use for shell work, not unrelated tasks merely running on Linux.

• security

  Handle requested security reviews, secure defaults, hardening, vulnerability reports, and sensitive-code ownership or bus-factor analysis. Includes Python, JS/TS, and Go guidance plus language-independent Git ownership tools. Use for security maintainers, orphaned sensitive code, and ownership maps;

• self-debug

  Diagnose failures inside Gestalt itself by correlating Mobile control-plane traces, relay persistence, Codex runtime records, agent topology, browser delivery, and sandbox behavior. Use when a Gestalt session stalls, loses a continuation or agent, reports a checkpoint or tool inconsistency, shows ba

• serena

  Navigate symbols, trace references, inspect diagnostics, and make structured code edits in the active workspace with verified managed Serena tools. Use only when the runtime supplies a working Codex-context Serena connection; retain native tools for unsupported languages.

• systematic-debugging

  Investigate bugs, test failures, build failures, performance regressions, and unexpected behavior before proposing a fix. Use when a reproducible symptom needs evidence, a root cause, a falsifiable hypothesis, and regression verification.

• verification-before-completion

  Match claims of complete, fixed, passing, verified, review-ready, or safe-to-ship work to current evidence. Use immediately before reporting a result or accepting delegated work.

• writing-skills

  Create or revise reusable agent skills, discovery metadata, bundled references, scripts, assets, and validation. Use when instructions must become concise, discoverable, specific, and reliable across stronger and lesser language models.

• xerj

  Use XERJ for bounded lexical queries during broad or repeated source discovery in fresh covered repositories. Preserve provenance and aliases, verify current source, and keep exact known-file lookup native.

MCP servers (1)

• gestalt-org-plan

  runs: node ./org-plan-mcp.mjs

Similar plugins