code-moniker
Use Code Moniker to onboard into an unfamiliar project, develop with indexed structural evidence, guard executable architectural memory after a change, or review relationships and change impact. Use it when callers, callees, coupling, ownership, project rules, or workspace-wide structure matter. Do not invoke it for routine Git, tests, exact-string lookup, or small known-file edits whose contract is already visible.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-26
- Stars
- ★ 3
- Forks
- 0
01 前言 Front Matter
- name
- code-moniker
- description
- Use Code Moniker to onboard into an unfamiliar project, develop with indexed structural evidence, guard executable architectural memory after a change, or review relationships and change impact. Use it when callers, callees, coupling, ownership, project rules, or workspace-wide structure matter. Do not invoke it for routine Git, tests, exact-string lookup, or small known-file edits whose contract is already visible.
02 资源文件 4
- 参考文档postures/develop.md2.2 KB
- 参考文档postures/guard.md2.2 KB
- 参考文档postures/onboard.md2.0 KB
- 参考文档postures/review.md2.0 KB
03 SKILL.md 正文
Code Moniker
Code Moniker combines project-authored architectural testimony with a symbolic
index. This skill routes agents to the product-owned learning surfaces instead
of duplicating their contracts. Rule and architecture concepts live in the
executable rules learn corpus; indexed request capabilities live in
query.describe and MCP tool schemas.
Choose a posture
Read only the posture that matches the current responsibility, then follow it before acting:
| Posture | Use it when | Playbook |
|---|---|---|
| Onboard | Entering an unfamiliar project or rebuilding project context | postures/onboard.md |
| Develop | Focusing structural exploration and implementing a change | postures/develop.md |
| Guard | Reassessing executable architectural memory after a change | postures/guard.md |
| Review | Evaluating architecture, relationships, risks, or change impact | postures/review.md |
Postures may follow one another. A typical change uses Onboard once, Develop for the implementation, and Guard after validation. Review remains read-only unless the user separately authorizes changes.
Learn from the product
The skill is a router, not a substitute for rules learn. Run the relevant
topic before first interpreting or changing that concept in the current task.
Do not skip it because the concept appears familiar, and do not run every topic
when only one is relevant. Re-run a topic when the Code Moniker version or the
relevant learn corpus changed.
| Need | Required command | What it teaches |
|---|---|---|
| Rule structure and expressions | code-moniker rules learn basics |
Rule blocks, ids, predicates, severity, messages, and rationale |
| Project vocabulary and testimony | code-moniker rules learn taxonomy |
Patterns, components, natural rule ids, aliases, and diagnostic interpretation |
| Language tags and rule namespaces | code-moniker rules learn languages |
Language-specific recipes, plus independent TS, TSX, JS, and JSX policy over a shared analysis ecosystem |
| Architectural locations | code-moniker rules learn paths |
Moniker path patterns and reusable aliases |
| Local architecture files | code-moniker rules learn fragments |
Fragments, view URIs, merging, and namespaced rule ids |
| Cross-symbol boundaries | code-moniker rules learn refs |
Reference rules for imports, calls, inheritance, annotations, and layers |
| Child collections | code-moniker rules learn collections |
Collection predicates and multiset operations over child symbols |
| Iterated symbol sets | code-moniker rules learn domains |
Domains, descendants, pairs, shapes, and quantifiers |
| Local structural measures | code-moniker rules learn metrics |
Named metrics bound to the current or iterated symbol |
| Distribution measures | code-moniker rules learn aggregates |
Numeric aggregates, dispersion, entropy, and mode |
| Moniker relationships | code-moniker rules learn relations |
Ancestor, descendant, and binding relation operators |
| Structural directives | code-moniker rules learn directives |
Layout, correlated-existence, and moniker-segment directives |
| Rule-set selection | code-moniker rules learn profiles |
Defaults, profiles, warning severity, and suppressions |
| Indexed Query DSL | code-moniker query 'query.describe' |
Live verbs, fields, defaults, constraints, pagination, projections, and preferred MCP tools |
Language, framework, architecture-pattern, and workspace recipes are discovered
from the executable catalog rather than duplicated in this router. Run
code-moniker rules learn for progressive human navigation or
code-moniker rules learn --format json for the complete machine-readable
inventory, then load the relevant name or alias with
code-moniker rules learn <topic>. Examples
include java, spring, java-qualified-types, react, javascript, jsx,
and sql.
If a required learn command is unavailable, report that limitation instead of reconstructing its semantics from this skill.
Shared operating contract
- Use Code Moniker only when structural or architectural evidence can change the answer. Prefer ordinary repository tools for direct local facts.
- Select one symbolic surface. Prefer available
code_moniker_*MCP tools; use the local CLI when MCP is unavailable. Hooks enforce write-time policy and do not replace exploration. - For advanced indexed reads, discover the live Query DSL before composing it:
use
code_moniker_querywithquery:"query.describe"through MCP, orcode-moniker query 'query.describe'through the CLI. To target an existing CLI index exactly, take its endpoint fromcode-moniker daemon listand passquery --daemon <ENDPOINT>; do not infer an endpoint or fall back to a filesystem command. - Verify workspace identity once before the first workspace-wide MCP read and again only after roots or the connection change.
- Start compact, narrow, and bounded. Page, broaden, or request source only when omitted evidence can change the answer.
- Reuse returned monikers and generated follow-up calls rather than guessing identities or reconstructing queries.
- Keep project testimony, indexed facts, architectural interpretation, and coverage limits distinct. Attribute a finding only to the surface exercised.
- Stop when the evidence is sufficient. Do not replay the same exploration through MCP, CLI, and direct daemon queries.
04 同仓库更多 Skills
code-moniker
暂无简介
code-moniker-smell-review
Review a repository with code-moniker's local and workspace check DSL for Fowler and Lanza-Marinescu code smells. Use when Codex needs to write, validate, or run project-specific warning-severity smell rules, distinguish executable local or indexed DSL checks from AST/history checks, triage check output, or plan follow-up evolutions for missing smell-detection operators.