Using Skill Engine
When the user invokes any engine workflow from a contextualizer directory and the appropriate sub-workflow needs to be selected based on the directory's setup state.
Install
Quick install
npx skills add https://github.com/nick-railsback/skill-enginenpx skills add nick-railsback/skill-engine --agent claude-codenpx skills add nick-railsback/skill-engine --agent cursornpx skills add nick-railsback/skill-engine --agent codexnpx skills add nick-railsback/skill-engine --agent opencodenpx skills add nick-railsback/skill-engine --agent github-copilotnpx skills add nick-railsback/skill-engine --agent windsurfMore install options
Shorthand — useful for multi-skill repos:
npx skills add nick-railsback/skill-engineManual — clone the repo and drop the folder into your agent's skills directory:
git clone https://github.com/nick-railsback/skill-engine.gitcp -r skill-engine ~/.claude/skills/Using Skill Engine
When the user invokes any engine workflow from a contextualizer directory and the appropriate sub-workflow needs to be selected based on the directory's setup state.
---
name: using-skill-engine
description: When the user invokes any engine workflow from a contextualizer directory and the appropriate sub-workflow needs to be selected based on the directory's setup state.
---
Using the skill engine
This is the entry-point skill. It detects whether the current directory holds a
contextualizer that has already been set up, and routes to the matching
workflow.
Routing
When invoked, do the following in order. A contextualizer is installed at.claude/skills/<slug>-context/; its research/.research-state.json is
the canonical setup-state marker.
From the project working directory (the parent of .claude/), locate the
contextualizer root:
ctx_roots=$(find .claude/skills -mindepth 1 -maxdepth 1 -type d -name '*-context' 2>/dev/null)
ctx_count=$(printf '%s
' "$ctx_roots" | grep -c .)
- No contextualizer installed.
ctx_count == 0⇒ this is a fresh
.claude/skills/*-context/ was found,
then hand off to the engine-bootstrap skill to scaffold from
templates.
- Contextualizer root present, state file absent or unparseable.
[ ! -f "$ctx_root/research/.research-state.json" ] OR
jq empty "$ctx_root/research/.research-state.json" exits non-zero ⇒
the state file is missing or corrupt. Surface a one-line diagnostic
naming the path and the parse error, then route to
engine-bootstrap so the maintainer can re-scaffold over the broken
substrate (the bootstrap workflow surfaces an existing-files warning
before overwriting).
- Multiple contextualizer roots present.
ctx_count > 1⇒ surface
- State file present and parses. Route to the workflow named in the user's
refresh— full freshness sweep across tracked resourcesnew-reference— register a new resource and create the referencediscover— goal-given scan that writes references for what mattersstatus— read-only freshness dashboardself-audit— read-only drift audit
The chapter doctrine in 03-engine.md enumerates six workflows
(REFRESH, SKILL, NEW, STATUS, DISCOVER, SELF-AUDIT). The plugin surface
ships five distinct slash-commands plus the router, the scaffolder,
and clean-cache (eight skills total): the chapter's SKILL workflow
— single-reference targeted update — is reachable via new-reference
with an existing reference named in scope, so the two collapse to one
plugin command. /skill-engine:clean-cache is invoked directly, not
routed through this entry-point skill.
If the user did not name a workflow, render the menu from the engine
chapter's "The menu" section and wait for the human to pick one.
Doctrine surfaces
The full activation protocol (engine doctor, reflections, rejection-log
preflight) and the menu live in the engine chapter 03-engine.md under## Activation and ## The menu (six workflows). The orchestrator the
maintainer pastes into a fresh Claude Code session is the contextualizer's
navigator skill at .claude/skills/<slug>-context/SKILL.md, which is
stamped from one of the navigator templates underengine-bootstrap-templates/ by the engine-bootstrap skill.
Routing in this revision is the binary present-or-absent check above; a richer
compatibility audit is deferred to a later revision.
---
Source: https://github.com/nick-railsback/skill-engine
Author: nick-railsback
Discovered via: skillsdirectory.com
Genre: development
SKILL.md source
---
name: Using Skill Engine
description: When the user invokes any engine workflow from a contextualizer directory and the appropriate sub-workflow needs to be selected based on the directory's setup state.
---
# Using Skill Engine
When the user invokes any engine workflow from a contextualizer directory and the appropriate sub-workflow needs to be selected based on the directory's setup state.
---
name: using-skill-engine
description: When the user invokes any engine workflow from a contextualizer directory and the appropriate sub-workflow needs to be selected based on the directory's setup state.
---
# Using the skill engine
This is the entry-point skill. It detects whether the current directory holds a
contextualizer that has already been set up, and routes to the matching
workflow.
## Routing
When invoked, do the following in order. A contextualizer is installed at
`.claude/skills/<slug>-context/`; its `research/.research-state.json` is
the canonical setup-state marker.
From the project working directory (the parent of `.claude/`), locate the
contextualizer root:
```bash
ctx_roots=$(find .claude/skills -mindepth 1 -maxdepth 1 -type d -name '*-context' 2>/dev/null)
ctx_count=$(printf '%s
' "$ctx_roots" | grep -c .)
```
1. **No contextualizer installed.** `ctx_count == 0` ⇒ this is a fresh
project with no contextualizer yet. Route to **engine-bootstrap**:
surface a one-line note that no `.claude/skills/*-context/` was found,
then hand off to the `engine-bootstrap` skill to scaffold from
templates.
2. **Contextualizer root present, state file absent or unparseable.**
`[ ! -f "$ctx_root/research/.research-state.json" ]` OR
`jq empty "$ctx_root/research/.research-state.json"` exits non-zero ⇒
the state file is missing or corrupt. Surface a one-line diagnostic
naming the path and the parse error, then route to
**engine-bootstrap** so the maintainer can re-scaffold over the broken
substrate (the bootstrap workflow surfaces an existing-files warning
before overwriting).
3. **Multiple contextualizer roots present.** `ctx_count > 1` ⇒ surface
the list and ask the user which contextualizer to operate on. Do not
guess.
4. **State file present and parses.** Route to the workflow named in the user's
invocation context. The five plugin-surfaced maintenance workflows are:
- `refresh` — full freshness sweep across tracked resources
- `new-reference` — register a new resource and create the reference
- `discover` — goal-given scan that writes references for what matters
- `status` — read-only freshness dashboard
- `self-audit` — read-only drift audit
The chapter doctrine in [`03-engine.md`](https://github.com/nick-railsback/skill-engine/blob/main/plugin/skill-engine/docs/03-engine.md) enumerates six workflows
(REFRESH, SKILL, NEW, STATUS, DISCOVER, SELF-AUDIT). The plugin surface
ships five distinct slash-commands plus the router, the scaffolder,
and `clean-cache` (eight skills total): the chapter's `SKILL` workflow
— single-reference targeted update — is reachable via `new-reference`
with an existing reference named in scope, so the two collapse to one
plugin command. `/skill-engine:clean-cache` is invoked directly, not
routed through this entry-point skill.
If the user did not name a workflow, render the menu from the engine
chapter's "The menu" section and wait for the human to pick one.
## Doctrine surfaces
The full activation protocol (engine doctor, reflections, rejection-log
preflight) and the menu live in the engine chapter [`03-engine.md`](https://github.com/nick-railsback/skill-engine/blob/main/plugin/skill-engine/docs/03-engine.md) under
`## Activation` and `## The menu (six workflows)`. The orchestrator the
maintainer pastes into a fresh Claude Code session is the contextualizer's
navigator skill at `.claude/skills/<slug>-context/SKILL.md`, which is
stamped from one of the navigator templates under
`engine-bootstrap-templates/` by the `engine-bootstrap` skill.
Routing in this revision is the binary present-or-absent check above; a richer
compatibility audit is deferred to a later revision.
---
**Source**: https://github.com/nick-railsback/skill-engine
**Author**: nick-railsback
**Discovered via**: skillsdirectory.com
**Genre**: development
Related skills 6
caveman
Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.
secure-linux-web-hosting
Use when setting up, hardening, or reviewing a cloud server for self-hosting, including DNS, SSH, firewalls, Nginx, static-site hosting, reverse-proxying an app, HTTPS with Let's Encrypt or ACME clients, safe HTTP-to-HTTPS redirects, or optional post-launch network tuning such as BBR.
readme-i18n
Use when the user wants to translate a repository README, make a repo multilingual, localize docs, add a language switcher, internationalize the README, or update localized README variants in a GitHub-style repository.
lark-shared
Use when first setting up lark-cli, running auth login, switching user/bot identity (--as), handling permission denied or scope errors, needing to update lark-cli, or seeing _notice in JSON output.
improve-codebase-architecture
Find deepening opportunities in a codebase, informed by the domain language in CONTEXT.md and the decisions in docs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable.
paper-context-resolver
Optional RigorPilot helper for README-first deep learning repo reproduction. Use only when the README and repository files leave a narrow reproduction-critical gap and the task is to resolve a specific paper detail such as dataset split, preprocessing, evaluation protocol, checkpoint mapping, or runtime assumption from primary paper sources while recording conflicts. Do not use for general paper summary, repo scanning, environment setup, command execution, title-only paper lookup, or replacin...