agentleFS
rakaarwakyrakaarwaky/agents-arwaky/skills/architechture/aes-layer/SKILL.md

aes-layer

Skill0 starsChanged 4 days agoMITView on GitHub

AES layer scaffolding for agent, capabilities, contract, root, surface, taxonomy, utility across Python, Rust, TypeScript, plus test-suite HOW-TOs under reference/test/. Create a new layer file or improve/audit an existing one.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 0 are distinctive, and 7 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 7

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Verify: lint-arwaky-cli scan <layer-path> → 0 violations.

    line 22
  2. Very specific:

    Every reference/<layer>/<language>.md HOW-TO must use identical H2 heading order so a contributor can jump between Python, Rust, and …

    line 28
  3. Very specific:

    Every language variant gets the Import Matrix table under ### Import rules: Allowed imports, Mandatory imports (AES202), Forbidden imports.

    line 40
  4. Very specific:

    Every capabilities HOW-TO gets a Section Contract table with an Enforced column naming the AES403 check (e.g.

    line 41
  5. Very specific:

    Every agent HOW-To gets an "Error and VO rules" block with Rule 1 "Never silently discard" through Rule 5 (primitive/bool/domain-field …

    line 42
  6. Very specific:

    No extra top-level ## sections between Review and Verify — fold Anti-Patterns, Dogfood, or any other extra section into Review as ### …

    line 45
  7. Very specific:

    Verify H2 order with grep -n "^## " reference/<layer>/<language>.md — output must match Rules → Template → Section Contract → Workflow → …

    line 47

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

aes-layer

Scaffold AES layer files — agent orchestrators, capabilities, contracts, composition roots, surface commands/UI, taxonomy VOs, and stateless utilities — across Python, Rust, and TypeScript. Each layer owns one domain: agent implements exactly one aggregate; capabilities implement one protocol; contract defines protocol ABCs; root wires containers and entries; surface maps input events to VOs; taxonomy holds shared VOs/constants; utility holds pure helpers. The layer chain reads contract → capabilities → agent, consumed by surface/root. One layer file implements one protocol/aggregate pair where applicable. Invariants read by more than one layer live in Global Rules below. The layer also provisions test-suite HOW-TOs under reference/test/ (python/rust/typescript), moved here from the former aes-testing-suite skill.

Rules

  • Argument shape: <path> [create|improve] [--dry-run]. <path> is a folder (crate/module/package) or a single file; missing mode defaults to improve.
  • Language auto-detection: resolve the path, detect python (.py) / rust (.rs) / typescript (.ts/.tsx/.js/.jsx/.mts/.cts) from files present; on mixed or ambiguous input, list the detected languages and proceed with the dominant one, naming it in the report.
  • Routing: dispatch to reference/<layer>/<language>.md based on the detected layer and language; that language HOW-TO owns the workflow steps.
  • Mode create: scaffold a new file from the language HOW-TO template section.
  • Mode improve: audit the named file/folder against the language HOW-TO rules, list findings, apply fixes.
  • Safety model: default (no --dry-run) = autonomous, no permission prompts, execute directly. --dry-run = read-only: produce the findings report and STOP before any write; the report is the deliverable. This bullet is the single source of truth for mode/safety semantics.
  • Report: every run ends with a report block — files touched (dry-run: files that WOULD be touched), language detected, mode, findings/fixes.
  • Verify: lint-arwaky-cli scan <layer-path> → 0 violations. Language compile is fallback only.
  • On-disk verify: a parallel agent or git reset can revert files without notice — after applying any edit, re-read the touched region (or grep the new content) before declaring the run done.
  • One fact, one home: the language HOW-TO carries its layer's suffix, imports, and structure — read it, do not restate here. Cite the linter, not this file, when pointing at a rule. Each bullet states a fact once; references point at the bullet.

Cross-language consistency (all 3 languages)

Every reference/<layer>/<language>.md HOW-TO must use identical H2 heading order so a contributor can jump between Python, Rust, and TypeScript variants without relearning structure:

## Rules
## Template
## Section Contract
## Workflow
## Review
## Verify
  • Workflow is its own ## heading — never nested under ## Rules as ###. The agent files set this standard; capabilities files must match it.
  • Every language variant gets the Import Matrix table under ### Import rules: Allowed imports, Mandatory imports (AES202), Forbidden imports. Do not drop the Mandatory row from any language file.
  • Every capabilities HOW-TO gets a Section Contract table with an Enforced column naming the AES403 check (e.g. CapabilityBlockMarkers, CapabilityNoImplementor) and the severity.
  • Every agent HOW-To gets an "Error and VO rules" block with Rule 1 "Never silently discard" through Rule 5 (primitive/bool/domain-field rules), and the AES405-11 ≥2 injected subsystems note in the constructor comment.
  • Rust template placeholders must use the standard <Name>, <VO>, <domain> tokens — never <name-feature>, <name-policy>, <name-store>, or similar per-role variants. A Rust file with mixed placeholder conventions is a defect, not a language variant.
  • Watch for ORCA_RICH_MD inline-html artifacts ([[ORCA_RICH_MD:...:inline-html:...]]) in Rust files — they are paste corruption, remove them and rewrite the line in plain markdown.
  • No extra top-level ## sections between Review and Verify — fold Anti-Patterns, Dogfood, or any other extra section into Review as ### subsections. A sibling ## Anti-Patterns or ## Dogfood breaks the cross-language order and is a defect.
  • lint-arwaky-cli scan --layer <layer> → 0 violations is the gate; language compile (tsc --noEmit, python -c "import ...", cargo check -p <crate>) is fallback only.
  • Verify H2 order with grep -n "^## " reference/<layer>/<language>.md — output must match Rules → Template → Section Contract → Workflow → Review → Verify exactly; no extra, no missing, no reordered.

Routing

FilePurpose
reference/agent/python.mdPython agent HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/agent/rust.mdRust agent HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/agent/typescript.mdTypeScript agent HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/capabilities/python.mdPython capabilities HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/capabilities/rust.mdRust capabilities HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/capabilities/typescript.mdTypeScript capabilities HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/contract/python.mdPython contract HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/contract/rust.mdRust contract HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/contract/typescript.mdTypeScript contract HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/root/python.mdPython root HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/root/rust.mdRust root HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/root/typescript.mdTypeScript root HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/surface/python.mdPython surface HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/surface/rust.mdRust surface HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/surface/typescript.mdTypeScript surface HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/taxonomy/python.mdPython taxonomy HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/taxonomy/rust.mdRust taxonomy HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/taxonomy/typescript.mdTypeScript taxonomy HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/utility/python.mdPython utility HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/utility/rust.mdRust utility HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/utility/typescript.mdTypeScript utility HOW-TO — both modes; rules, template, section contract, workflow steps, review, verify
reference/test/python.mdPython test-suite HOW-TO — prefixes, run commands, coverage, verify
reference/test/rust.mdRust test-suite HOW-TO — prefixes, nextest, criterion, verify
reference/test/typescript.mdTypeScript test-suite HOW-TO — prefixes, vitest, coverage, verify

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.