core

Plugin core from marketplace rene-roscher-skills in Rene-Roscher/skills. Commands read from the repo's manifest — 5 days ago.

Skills (21)

• agent-delegation

  Decides when an orchestrating agent hands work to subagents, how many run at once, what the brief holds and which model tier each gets by risk and difficulty: small for lookups, balanced for routine work, strong for planning and business-critical code, top tier only for the hardest problems. Use whe

• agent-toolbelt

  Picks the cheapest tool that answers a question and keeps the working context small: file-name and text search before reading, jq for JSON, structural search, git history search, gh for GitHub objects, type checker or language server for symbols, docs index tool, parallel calls, scoped subagents. Us

• bug-investigation

  Takes a defect from report to regression test without guessing: facts and blast radius first, safe reproduction, a test that fails without the fix, a root-cause fix, closure after a verified deploy. Use when given a stack trace, error-tracker link or issue ID, 'why does X crash', 'error in productio

• commit-hygiene

  Writes commits, branch history and pull request text the way a careful colleague would: one logical change, imperative subject, a body that says why, no agent attribution, session links, emoji or filler. Use when committing, amending, squashing, writing a PR description or release notes, or setting

• evidence-before-claims

  Checks version-dependent and external facts against a source before stating them: APIs, flags, defaults, standards, statutes, subagent findings, and the status, authorship and security claims a README or spec makes before a push. Labels each claim verified, reported or inferred. Use when an answer d

• executor-spec

  Lets a strong model plan a task completely and hand only the execution to a weaker or open-weight model: a self-contained spec with every decision made, exact edits and a check per step, acceptance tests written and hashed first, a file allowlist and a stop-and-ask protocol, then lint, a cold read a

• feature-delivery

  Implements an approved feature plan and closes it with evidence: an entry gate on the plan, stop and ask when a plan assumption breaks, one write path for UI and API, a definition of done met with evidence, tested kept apart from verified. Use when told 'implement the plan', 'build it now', 'go ahea

• feature-planning

  Plans a feature end to end before any code and stops at a plan gate until the owner approves: business goal and problem, depth by stakes, a cost model for cost-intensive features, UX states, data model, permissions, tenant isolation, legal triggers, performance budget, failure modes, rollout, open d

• glossary

  Resolves short names and abbreviations from the world of a German or EU software business (hosting, panels, payments, accounting, DSGVO and tax terms, analytics, mail, DNS, ops tools, SaaS metrics) before acting on them. Project glossary first, then bundled references, then the vendor site, then a q

• goal-focus

  Keeps an agent on the stated goal: restates the goal and its done criterion (for autonomous work also the business and feature goal, weighed against UX, cost and security), keeps a short task list, checks each step against the goal, notes side findings instead of fixing them, timeboxes investigation

• know-your-limits

  Working rules for a model below the strongest tier (small or balanced class, or one that cannot name its tier): nothing from memory, names looked up before use, one small step then a check, shipped checkers decide, procedures followed literally, hard stop-and-hand-off triggers, calibrated reports. U

• learning-loop

  Stops an agent from repeating a mistake. After a correction or a non-obvious trap it records a lesson (rule, Why, How to apply, links) in the project's durable memory, checks lessons before acting, promotes a recurring lesson into a test, lint rule, hook or CI gate, and fixes or deletes stale ones.

• maintainable-code

  Keeps new code easy to change and review: seams and adapters so it is testable, search the repo before writing, follow neighbouring conventions, duplication before the wrong abstraction, small public surface, explicit errors, dead code and flags removed, small changes, a checklist before adding a de

• performance-gate

  Gates a feature with data or traffic impact before handover: query budget per request measured with a frozen clock, N+1, indexes tied to their query, EXPLAIN on the real engine, tagged cache invalidation, dashboards loaded per widget and cached per tenant, queue and broadcast fan-out, paid-message c

• project-inception

  Sets up docs/ before any feature code: product, business and per-area specs, ADRs, plans, project facts and an AGENTS.md read-first table, each written with the skill that owns the area, gated on owner approval. Adopts undocumented repos by deriving specs from code and listing drift. Use when starti

• project-scaffold

  Creates the code skeleton of an approved project from each stack's official starter, only on request and only for the components the approved docs name: a Laravel-only project gets the Laravel starter and nothing else. Pinned versions, flags instead of prompts, an untouched first commit, the starter

• safe-actions

  Classifies every action with effects outside the working tree before it runs (allowed, approval, never) and names a safe substitute. Use when about to run kubectl, terraform apply, ssh, a deploy, a non-local database write, a test mail, SMS or payment, git push, a release or making a repository publ

• security-assurance

  Proves each security property a product relies on or claims instead of trusting labels: asset, attacker and mechanism in its own words, then checks how each technique is used (AEAD and nonces, keys and rotation, password hashing, CSPRNG tokens, constant-time compares, token algorithms, TLS verificat

• stack-bootstrap

  Detects a repository's stack from its files (composer.json, go.mod, package.json, Dockerfile, workflows, Terraform, Helm, Kustomize, Ansible) and enables the matching plugins of this marketplace plus a curated third-party set, per project. Use when setting up the agent in a new or cloned repo, or af

• test-strategy

  Decides what to test and how tests carry weight: critical paths (money, auth, tenant isolation, deletion, permissions) first, each test seen failing, assertions on outcomes not mocks, no weakened assertions, flaky tests as bugs, coverage gates on critical modules, spike load tests. Use when planning

• verify-before-done

  Separates tested from verified, and merged from live, before reporting done: a runtime gate on the running code, regression tests that fail without the fix, test-versus-production gaps, no vacuous checks. Use when about to say done, fixed, works, shipped or live, when a fix has no effect, before clo

Similar plugins