How do I install the Context Warm Related skill?

Run npx skillmds@latest add jokerman89/context-warm-related in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Context Warm Related safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Context Warm Related?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Context Warm Related free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Context Warm Related?

jokerman89 (@jokerman89) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Context Warm Related by jokerman89 — Agent Skill

Name: Context Warm Related
Author: jokerman89

File contents

You are the context-warm-related skill — topic-heuristic loading.

What this skill does

Given a topic (e.g., "rate limiter", "auth flow", "retry pattern"), heuristically finds the top N most-relevant files across cwd + ~/.lintel/scaffolding/ + .claude/engineering/design-archive/ and loads them via /li:context-warm.

When to use

Operator says "warm context with everything related to X"
Pre-PLAN when DISCOVER identified topic-relevant files
Cross-repo topic understanding (e.g., compare current repo's X with scaffolding template's X)

When NOT to use

Specific files known — use /li:context-warm <paths> directly
Whole-repo context needed — too broad for heuristic, use targeted patterns

Workflow

Step 1 — Topic + scope

topic="$1"
scope="${2:-cwd}"  # cwd | repo | lintel-home | all
limit="${3:-10}"

Step 2 — Heuristic search

Multi-source search:

Filename match: find . -iname "*<topic>*"
Content match: grep -rli "<topic>" --include='*.md' --include='*.ts' --include='*.py' --include='*.go'
Frontmatter match (for skills/agents): grep -li "<topic>" skills/*/SKILL.md agents/*/*.md

Score files:

3 points: filename match
2 points: ≥5 content matches
1 point: 1-4 content matches

Top N by score.

Step 3 — Surface candidate list

CANDIDATE FILES (related to "<topic>") — top <N> by relevance:

| Score | File | Match type |
|---|---|---|
| 5 | .claude/engineering/design-archive/lintel-v3-plan.md | filename + 12 content |
| 4 | src/lib/rate-limiter.ts | filename + 8 content |
| 3 | .claude/engineering/design-archive/CONTEXT-ENGINE.md | 7 content |
| ... | | |

Estimated tokens to load all <N>: ~<X>k

Load all? (Y / select subset / cancel)

Step 4 — Delegate to context-warm

If operator selects all or subset:

/li:context-warm <space-separated-paths>

Step 5 — 00-state.md append

event: context_warm_related
ts: <timestamp>
topic: <topic>
candidates_surfaced: <N>
loaded: <count>
tokens_added: <approx>

Integration

Reads codebase + delegates to /li:context-warm. Writes to budget tracking via that skill.

Anti-patterns

Loading 50+ files because "related" is broad — cap at N=10 default, operator can extend
Skipping confirmation — heuristic match isn't always right
Using broad topics like "api" — too many matches; suggest narrower

---name: context-warm-relateddescription: Heuristic context warm — search codebase for files related to a topic, load top N most-relevant.---You are the context-warm-related skill — topic-heuristic loading.## What this skill doesGiven a topic (e.g., "rate limiter", "auth flow", "retry pattern"), heuristically finds the top N most-relevant files across cwd + `~/.lintel/scaffolding/` + `.claude/engineering/design-archive/` and loads them via `/li:context-warm`.## When to use- Operator says "warm context with everything related to X"- Pre-PLAN when DISCOVER identified topic-relevant files- Cross-repo topic understanding (e.g., compare current repo's X with scaffolding template's X)## When NOT to use- Specific files known — use `/li:context-warm <paths>` directly- Whole-repo context needed — too broad for heuristic, use targeted patterns## Workflow### Step 1 — Topic + scope```bashtopic="$1"scope="${2:-cwd}" # cwd | repo | lintel-home | alllimit="${3:-10}"```### Step 2 — Heuristic searchMulti-source search:- Filename match: `find . -iname "*<topic>*"`- Content match: `grep -rli "<topic>" --include='*.md' --include='*.ts' --include='*.py' --include='*.go'`- Frontmatter match (for skills/agents): `grep -li "<topic>" skills/*/SKILL.md agents/*/*.md`Score files:- 3 points: filename match- 2 points: ≥5 content matches- 1 point: 1-4 content matchesTop N by score.### Step 3 — Surface candidate list```CANDIDATE FILES (related to "<topic>") — top <N> by relevance:| Score | File | Match type ||---|---|---|| 5 | .claude/engineering/design-archive/lintel-v3-plan.md | filename + 12 content || 4 | src/lib/rate-limiter.ts | filename + 8 content || 3 | .claude/engineering/design-archive/CONTEXT-ENGINE.md | 7 content || ... | | |Estimated tokens to load all <N>: ~<X>kLoad all? (Y / select subset / cancel)```### Step 4 — Delegate to context-warmIf operator selects all or subset:```bash/li:context-warm <space-separated-paths>```### Step 5 — 00-state.md append```yamlevent: context_warm_relatedts: <timestamp>topic: <topic>candidates_surfaced: <N>loaded: <count>tokens_added: <approx>```## IntegrationReads codebase + delegates to `/li:context-warm`. Writes to budget tracking via that skill.## Anti-patterns- **Loading 50+ files because "related" is broad** — cap at N=10 default, operator can extend- **Skipping confirmation** — heuristic match isn't always right- **Using broad topics like "api"** — too many matches; suggest narrower

Frequently asked questions

Context Warm Related