Agentic Auto-Improve

Plugin agentic-auto-improve from marketplace auto-improve in lneninger/claude-agentic-auto-improve. Commands read from the repo's manifest — today.

Skills (34)

• advance

  Move a contract forward by exactly one step. Asks the script what the next move is, then takes it: start the sub-tasks that are ready, escalate a failure to the architect, land a finished contract that has a parent branch (a background script, no question asked) or hand one without a parent branch t

• angular-ui-rules

  The short Angular, Angular Material and Tailwind rule list, keyed by Rule ID, that the Angular agents preload. Use when: writing or reviewing Angular components, templates, state services or styles; or when an agent must cite a Rule ID. Points at the full rule documents installed by the angular-tail

• application-design

  The data-dense product-screen layer of the design method: density, the trading typography scale, metric grids, status badges, metric cards, loading skeletons, empty states, real-time value flashes, live regions and the sidebar-plus-content app shell, on top of the design foundation. Use when: design

• auto-improve-angular-tailwind-install

  Install the angular-tailwind add-on of the agentic-auto-improve plugin in this project: four rule documents (Angular code, Angular Material, Tailwind styling, UI patterns) copied into .claude/references/angular-tailwind/ and listed in the project profile's review.documents slot, plus, on request, th

• auto-improve-angular-tailwind-uninstall

  Remove the angular-tailwind add-on of the agentic-auto-improve plugin from this project: the four rule documents under .claude/references/angular-tailwind/ and their entries in the project profile's review.documents slot, and any copy of the retired prompt advisor hook (a settings.json row and a fil

• auto-improve-codegraph-install

  Install or upgrade the codegraph add-on of the agentic-auto-improve plugin in this project: one hook file with two rows that blocks a source Read, Grep or Glob until a CodeGraph tool has run in the same turn. Needs the CodeGraph MCP server; stops and says so plainly when it is not configured. Previe

• auto-improve-codegraph-uninstall

  Remove the codegraph add-on of the agentic-auto-improve plugin from this project: the settings.json rows of the code-search-first guard (and of the retired tracker and reset hooks, when an older install left them), then the hook files. Previews every row and file, asks for one confirmation, then wri

• auto-improve-database-guard-install

  Tell the database guard of the agentic-auto-improve plugin which databases in this project are protected: finds the database names this project's tracked configuration files use, shows each with the file and line that names it, asks which to protect (and whether a Testing-environment database exists

• auto-improve-finish-install

  Finish installing the agentic-auto-improve plugin in this project: fill in the project profile (Quick start step 4), propose the area map, the pull request module tags and the project-name tokens (Quick start step 3), and check or register the hooks (Quick start step 5). Looks at the repository, sho

• auto-improve-strict-shell-gate-install

  Install the strict-shell-gate add-on of the agentic-auto-improve plugin in this project: a hook that reads each shell command before it runs and blocks a write into a source file that no approved contract covers. Says plainly what it gets wrong. Previews the files and the settings.json change, asks

• auto-improve-strict-shell-gate-uninstall

  Remove the strict-shell-gate add-on of the agentic-auto-improve plugin from this project: the settings.json row of the bash-gate hook (PreToolUse on Bash), then the file .claude/hooks/bash-gate.py. Previews the row and the file, asks for one confirmation, then writes: the row first with a backup, th

• cloud-engineering

  Cloud and infrastructure knowledge for design, review, non-mutating validation and troubleshooting across Azure, AWS and Google Cloud, with Terraform, Bicep, Docker and Kubernetes. Reads cloud.providers and cloud.iac from the project profile and opens only the matching references. Never runs a live-

• contract-accuracy

  Read-only inspector for the contract-critic accuracy-skip system. Prints the per-area clean_streak, skip_eligible flag, and recent history. Use when: user says /contract-accuracy, 'show accuracy', 'which areas can skip critic', 'why did /design-first skip the critic on X', or wants to audit the syst

• critique-now

  Manually run contract-critic against an already-approved (or skipped) contract. Use when: user says /critique-now <path>, 'run critic anyway', 'audit this contract', 'I want a second opinion on this approved design', or wants to verify the accuracy-skip system was right to skip a previous contract.

• cross-impact

  Run the cross-area scanner against a concept contract or a list of files and print the top-3 adjacent areas with score breakdowns. Use when: the user types `/cross-impact <contract-path>` or `/cross-impact <comma-separated-files>`, or asks "what other areas would this contract affect?", "what's adja

• debug

  Systematic 4-phase debugging protocol — REPRODUCE → ISOLATE → HYPOTHESIZE → VERIFY before any fix is written. Forces root-cause analysis instead of symptom patching. Writes findings to .claude/debug-notes-<slug>.md. Use when: user says /debug, 'debug this', 'why is X failing', 'fix this bug', or whe

• design-first

  Kick off the Data-First Engineering Protocol for any non-trivial task. Classifies the task, invokes the data-architect to produce a concept contract at .claude/concepts/, resolves Open Questions with the user, and hands off to implementer agents with the contract as their authoritative input. Use wh

• flow

  Run the whole contract-driven delivery chain end to end in one invocation: intake, design contract, test-first implementation, adversarial review, verification, commit and pull request. Drives the working agent chain (/task, /design-first, /tdd-first, the reviewer agents, /verify-before-done, /git-c

• git-commit

  Commit staged changes safely. Runs /verify-before-done first, lets the user choose which files to stage, drafts a project-style commit message (subject + Files / Verification / Contract / Status sections + Co-Authored-By trailer), and commits via HEREDOC. Halts on verification failure, suspected sec

• journal-add

  Append a lesson-learned entry to the journal (.claude/registries/JOURNAL.md root or a .claude/registries/JOURNAL/<domain>.md file) with registry_append.py. Use when: the user says /journal-add, 'log a lesson', 'add to the journal', 'record this for next time', or after a production incident / debugg

• list-contracts

  Show every concept contract under .claude/concepts/ grouped by project and status. Highlights contracts needing user action (drafts with unresolved Open Questions, approved contracts not yet implemented). Use when: user says /list-contracts, 'show my contracts', 'what's pending', or after finishing

• north-star-review

  Run the north-star review engine against every active thought in the project. The engine reads each thought's anti-patterns, greps the live codebase for hits, and emits a ranked list of suggested next-step contracts. Optionally updates the `Current gap` and `Suggested next steps` sections of each th

• north-star

  Capture an aspirational direction-setting thought as a north-star file at .claude/north-stars/<YYYY-MM-DD>-<slug>.md. North-stars are the WHAT-WE-WANT register; distinct from JOURNAL.md (lessons from what already happened) and contracts (concrete tasks). Read by data-architect Step 0.5 on every /des

• plan-questions

  Question-quality rubric for AskUserQuestion calls during /plan mode. Triggers when the developer says /plan-questions, "ask better questions", "that question was confusing", or as a mid-session reset. Invoke it by name; the plugin no longer injects it automatically (the plan-question advisor hook wa

• pr-merged

  The closing phase of one sub-task: once its pull request has merged, record the completion and work out which pending sub-tasks that releases. Verifies the merge by asking GitHub rather than trusting the claim, writes a completion record stamped verified: github, recomputes the released set, and han

• promote-ui-rule

  Promote a recurring UI mistake from the journal (JOURNAL.md and JOURNAL/*.md) into an automated guard rule in architecture-guard.rules.json. The promotion is gated: requires 3+ journal entries tagged `ui-evolution + recurring-mistake` matching the same pattern. Use when: the user says `/promote-ui-r

• ship

  Turn a finished work item into a pull request that GitHub actually links back to its issue. Reads the Work Item Brief for the branch, worktree, contract and issue id, runs /verify-before-done, composes the PR title from the modules the diff really touches, pushes from the tree that holds the commits

• sql-server-patterns

  Inject SQL Server schema-design knowledge into a task before the schema is written — partitioning, hierarchy modelling, SARGability, compression and temporal tables — by routing to .claude/templates/tsql-patterns.md rather than restating it. Distinct from migration-safety-reviewer (will this migrati

• task

  Intake a work item — a GitHub issue (this project's tracker), or a Jira/ADO item, or a free-text request — and turn it into a normalized Work Item Brief at .claude/work-items/, with acceptance criteria, size classification, and an issue-linked branch cut inside a flat worktree of its own (every chan

• tdd-first

  Enforce a RED→GREEN→REFACTOR cycle for a feature. Writes failing tests BEFORE implementation, runs the test suite to confirm the red state, then invokes the implementer agent with the failing tests as authoritative input, then re-runs to confirm green. Use when: user says /tdd-first, 'do this test-f

• unit-testing

  How to write, run and report unit and integration tests in this repository: the test conventions for .NET (xUnit) and Angular (Vitest), the RED-first workflow, the rule that every resource a test run creates is removed afterwards whether the run passes or fails, and the format of the detailed result

• validate-registries

  Scan the registries (.claude/registries/MECHANISMS.md with MECHANISMS/*.md, and VOCABULARY.md with VOCABULARY/*.md) for dead file references, and with --check-structure lint their layout (bodiless entries, duplicate titles, project content in the Universal root, digest freshness). Reports any backti

• verify-before-done

  Run the verification gauntlet before claiming a task is complete. Executes build, tests, migration check, generated-model drift check, and HANDOFF completeness check. Blocks on any failure. Use when: an agent or the user is about to declare Status=complete, before committing, before opening a PR, or

• website-design

  The design foundation for any page of a project: theme polarity per application, the Material-to-Tailwind @theme token bridge, the inline-style ban, WCAG 2.1 AA contrast, keyboard and ARIA rules, purposeful motion and the never-do list. Use when: designing or restyling a website page, a marketing or

Subagents (15)

• angular-senior-dev

  Use for any Angular 21 frontend work in this solution — standalone components, signal state services, HTTP services, route guards, interceptors, routing. Use proactively when a task touches any application root named by the `frontend.roots` slot of `.claude/project-profile.md`. Implements production

• api-contract-reviewer

  Use this agent after any change to a .NET controller, request/response DTO, enum in an API payload, or an Angular core model. Detects breaking API contract drift between backend and frontend — property renames, enum changes, nullability shifts, route moves, ApiResponse envelope changes, status code

• cloud-engineer

  Use for cloud and infrastructure work on Azure, AWS or Google Cloud with Terraform, Bicep, Docker or Kubernetes. Covers design, review, non-mutating validation and troubleshooting under the `cloud.roots` slot of the project profile. Active only when `cloud.providers` is set. Never runs a live-changi

• contract-critic

  Use this agent during the Data-First Engineering Protocol AFTER the data-architect drafts a concept contract and BEFORE the user sees it for approval. Adversarially reviews the draft against MECHANISMS.md / VOCABULARY.md / INTEGRATION.md / JOURNAL.md and appends a `## Critique` section to the contra

• data-architect

  Use this agent FIRST on any non-trivial task. It designs the data and mechanism model BEFORE any implementation begins, reads MECHANISMS.md and VOCABULARY.md, and produces a concept contract at .claude/concepts/ that becomes the authoritative input for implementer agents. This agent NEVER writes pro

• dotnet-backend-architect

  Use for any C# / .NET 9 backend work in this solution — controllers, services, repositories, entities, EF Core configurations and migrations, DI wiring, Quartz jobs, and SignalR. Use proactively when a task touches any root named by the `backend.roots` slot of `.claude/project-profile.md`. Implement

• fullstack-code-reviewer

  Use this agent after any meaningful Angular or C# code has been written, modified, or refactored. This includes new features, bug fixes, refactors, and API changes. It performs adversarial review focused on correctness, security, performance, and alignment with project conventions.\n\nExamples:\n- A

• migration-safety-reviewer

  Use this agent after every `dotnet ef migrations add` and before any migration lands in a commit. Reviews the generated Up/Down against existing prod data shape — backfill strategy, NOT NULL adds, type compatibility, rename vs drop+add, online index adds, transaction safety, and reversibility. Disti

• python-ai-developer

  Use this agent when the user needs help with Python AI/ML model development, training, fine-tuning, architecture design, data pipelines, MLOps, or deployment. This covers deep learning (PyTorch, TensorFlow, JAX), LLM development (pretraining, alignment, RLHF, DPO), computer vision, NLP, multimodal m

• registry-scout

  Use this agent when an expensive agent needs to know which registry entries are relevant to a task, without reading the registries in full. It reads the generated digest at .claude/.scout-cache/registry-digest.md — a complete index of every mechanism, vocabulary term, journal lesson and integration

• security-auditor

  Use this agent after any change to auth, credentials, encryption, or user-data boundaries. Performs focused adversarial security review — JWT config, PBKDF2, encrypted third-party credentials, password reset flow, user-ownership enforcement, SignalR hub auth, CORS, rate limiting, and OWASP Top 10. D

• senior-test-engineer

  Use to author RED (failing) tests first in the TDD cycle — xUnit + Moq + FluentAssertions for .NET, Vitest + TestBed for Angular. Use proactively at the start of any non-trivial change before implementation. Writes tests ONLY; never writes production/implementation code. Follows the unit-testing ski

• sql-performance-reviewer

  Use this agent after any change to EF Core queries, repositories, or indexes — before the change ships. Focused on query plans, index coverage, N+1 patterns, client-side evaluation, tracking misuse, and hot-path cost. Distinct from schema-design work (handled via the `/sql-server-patterns` skill + `

• test-strategy-critic

  Use this agent after senior-test-engineer produces a test suite, or when a test suite is failing intermittently, or when coverage is high but bugs keep escaping. Reviews test *strategy*, not test syntax — what's worth testing, where mocks hide real bugs, and where integration boundaries actually mat

• ui-ux-designer

  Use this agent when designing or improving the visual presentation, layout, component aesthetics, theming, or user experience of this project's Angular frontend. This includes creating new UI layouts, improving data density, implementing dark theme, designing reusable visual components, ensuring acc

Similar plugins