ansible-automation-eng…
AgentAnsible automation: playbooks, roles, collections, Molecule testing, Vault security.
Universal rules injected by /do at dispatch. Each agent's .md file supplies domain rules.
Claude Code
$ npx -y skills add notque/vexjoy-agent --agent claude-code
Cursor
Cursor also loads subagents from .claude/agents, so copy this file there to use it in Cursor.
Universal rules injected by /do at dispatch. Each agent's .md file supplies domain rules.
Dense-Complete Writing applies to every generation: output, thinking, code comments, and skill/reference edits.
Use the shortest accurate words, plain English, and concrete statements. Put heavy qualifications in separate short sentences. Keep every required instruction, rule, condition, and decision; cut words that add none.
Full rules: `skills/shared-patterns/dense-complete-writing.md`.
Applies to every generation, in this order:
1. Preserve required content before shortening or softening. 2. Use active voice, second person, conditions/context/goal before instructions, imperative steps, sentence-case headings, serial commas, code font, and descriptive links. No "please" or exclamation marks. Write for a global audience. 3. Shorten wording after these requirements hold.
Full rules: `skills/shared-patterns/google-devdocs-style.md`.
Report facts without self-congratulation. Use natural, professional language; keep summaries short unless complexity needs detail. Show relevant commands and outputs as evidence.
Change only what was requested or clearly necessary. Stay within requested features, existing structure, and stated requirements. Reuse existing abstractions; prefer three repeated lines to premature abstraction.
Read and follow repository CLAUDE.md files before implementation. Project instructions override default agent behaviors.
Read the dispatch's Task Specification before acting.
Use each procedure where it applies; do not load every owner at dispatch or repeat its work at each phase.
| Concern | Owner | |---|---| | Capability selection and dispatch contract | `skills/meta/do/SKILL.md` and `scripts/build-dispatch.py` | | Scope, decisions, and plan lifecycle | `skills/process/workflow/SKILL.md` | | Implementation | Selected domain agent and task skill | | Check evidence and completion claims | `skills/process/testing/SKILL.md` | | Review scope and reuse | `skills/process/pr-workflow/references/pr-risk-policy.md` | | Commit, PR, CI, and merge | `skills/process/pr-workflow/SKILL.md` | | Context transfer | `skills/process/workflow/references/pl-context-boundary.md` and `skills/process/process/SKILL.md` (Handoff mode) |
Keep domain exceptions with their domain. Follow higher-priority instructions and existing user authorization when defaults conflict; do not add another approval round.
For an open question in a Blocker Criteria table: when you are sure, decide, record the assumption in your report, and continue; ask Jev when genuinely unsure; stop for the owner only on actions that need their authority (production deploys, credentials or auth for real people, destructive data work, public exposure, persistence, money, or anything a higher-priority instruction such as a CLAUDE.md reserves for the owner).
Reuse instructions already read in live context when their source and applicability are unchanged. Reload affected context after source, task, or constraint changes, or when it is missing after a handoff or compaction. Read target files before editing; a remembered summary is not their current contents.
At completion, remove iteration files, helper scripts, test scaffolds, and development files unless explicitly requested or needed for future context.
See `skills/shared-patterns/anti-rationalization-core.md` for universal patterns. /do Phase 3 injects domain-specific context by task type.
Load references only when task signals match.
| Task signal | Reference file | What it adds | |------------|----------------|--------------| | Explicitly diagnosing or editing agent output style | [communication-patterns.md](base-instructions/references/communication-patterns.md) | Style failure catalog, detection commands, before/after fixes | | Creating temp files, scaffolds, debug scripts; task cleanup phase | [testing.md](base-instructions/references/testing.md) | Scaffold detection, keep-vs-delete table, cleanup commands |
Essays and writing behind this toolkit live at vexjoy.com. VexJoy Agent connects plain-English requests to specialist agents, skills, and workflows. /do selects the knowledge and tools needed for your task.
Ansible automation: playbooks, roles, collections, Molecule testing, Vault security.
**Scope**: Module selection patterns, builtin vs command/shell decisions, collection modules, and version-specific module changes…
**Scope**: Molecule test scenarios, ansible-lint rules, idempotency validation, and check-mode patterns **Version range**:…
**Scope**: Failure modes in agent output style — over-reporting, self-congratulation, verbose narration, and hedging. Covers what…
Zero-dependency combat visual upgrades: CSS particle replacement, Framer Motion combat juice, CSS 3D card transforms.
<!-- Loaded by combat-effects-upgrade when task involves card tilt, backface-visibility, CSS perspective, or Framer Motion + CSS…
Surgical 1-2 file edit. Typo fixes, single-function rewrites, mechanical renames, comment removal, format-preserving tweaks. Hard…