load-context
Skill4 starsChanged 6 months agoMITView on GitHub
Load existing context files into the conversation for a feature branch. Use when user says "load context", "read context", "get context", or wants to resume work on a branch where /build-context was already run.
We split this skill into the rules it gives an agent
8 rules. 0 are common, 4 are distinctive, and 4 are very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
4
Distinctive
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
This is the complement to
/build-context: Use /build-context to create context files. - Distinctive: line 20
If no context files found:
- Distinctive: line 26
Read files in this order for optimal context placement:
- Very specific: line 32
If the tiered files don't exist (legacy single-file format), read whatever .md files are present.
- Very specific: line 38
api-context.md: check generated_at — this is cached and rarely stale
- Very specific: line 39
project-context.md: check generated_at — if older than 1 day, warn the user:
- Distinctive: line 50
Verify each file exists (skip missing with a note)
- Distinctive: line 51
Read all existing critical and important files using parallel agents if there are many files (>10)
“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
<!-- Generated from .agents/ by scripts/sync-ai-workspace.mjs. Do not edit directly. -->Load Context
Load pre-built context files into the current conversation so you can immediately proceed with planning or implementation — no research phase needed.
This is the complement to /build-context. Use /build-context to create context files. Use /load-context to reload them in a new conversation. Use /update-context to update stale project context without re-extracting API docs.
Instructions
Step 1: Detect Branch and Find Context Files
- Get the current git branch:
git branch --show-current - Glob for context files:
contexts/{branch-name}/**/*.md - If no context files found:
- Tell the user: "No context files found for branch
{branch-name}. Run/build-contextfirst." - Stop here.
- Tell the user: "No context files found for branch
Step 2: Read Context Files (tier-aware)
Read files in this order for optimal context placement:
context.md(Tier 3 — combined summary) — read first for orientationapi-context.md(Tier 1 — cached API reference) — read second for provider knowledgeproject-context.md(Tier 2 — dynamic project state) — read third for current state
If the tiered files don't exist (legacy single-file format), read whatever .md files are present.
Step 3: Check Freshness
Parse the YAML front matter from each tier file:
- api-context.md: check
generated_at— this is cached and rarely stale - project-context.md: check
generated_at— if older than 1 day, warn the user:⚠ Project context was built on {date}. Plans, reviews, and code may have changed. Run `/update-context` to update it (~1 min) or `/build-context` for a full rebuild.
Step 4: Read Must-Read Files
- From the context files, find the Must-Read Files section
- Extract file paths from Critical and Important tables (skip Reference)
- Deduplicate across all context files
- Verify each file exists (skip missing with a note)
- Read all existing critical and important files using parallel agents if there are many files (>10)
Step 5: Summary
Context loaded for branch `{branch-name}`.
Tiers:
- API context: {generated_at} {(N days old)}
- Project context: {generated_at} {(N days old)} {⚠ stale if > 1 day}
Context files read: {N}
Must-read files read: {N} critical, {N} important ({N} skipped — not found)
Critical path:
1. {from context file}
2. ...
Plan status:
- Completed: {summary}
- In progress: {summary}
Reference files available (not loaded): {N}
Ready for the next task.Critical Rules
- Never modify context files — this skill is read-only
- Read in tier order — context.md first (orientation), then api-context.md (reference), then project-context.md (current state)
- Skip reference-priority files — only read critical and important; mention reference files are available
- Warn on stale project context — if project-context.md is older than 1 day, suggest
/update-context - Deduplicate across files — if multiple files list the same must-read path, read it once
- Fail fast if no context exists — suggest
/build-context
Reference
.agents/skills/build-context/SKILL.md— creates context files (tiered).agents/skills/update-context/SKILL.md— lightweight update of project context
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- vishalmdi/ai-native-pm-os★ 102same name
- igmarin/rails-agent-skills★ 26same name
- ethanaubuchon/dossier-tradecraft★ 3same name
- nexus-a1/claude-skills★ 1same name
- rustimacc/context-ledger-skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository61 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .cursor/rules/ai-workspace-projections.mdcCursor rule
- .cursor/rules/mcp-tool-configs.mdcCursor rule
- .cursor/rules/node-typescript.mdcCursor rule
- .cursor/rules/runstore-listruns-cache.mdcCursor rule
- .cursor/rules/worker-posture.mdcCursor rule
- .cursor/rules/worker-safety.mdcCursor rule
- await-pr-checksSkill
- commitSkill
- create-agentSkill
- create-issueSkill
- create-mcpSkill
- create-planSkill
- create-prSkill
- create-ruleSkill
- create-skillSkill
- create-test-planSkill
- implement-planSkill
- load-contextSkill
- mergeprepSkill
- orchestrate-create-planSkill
- orchestrate-create-test-planSkill
- orchestrate-implement-planSkill
- orchestrate-resolve-pr-commentsSkill
- orchestrate-reviewSkill
- publish-releaseSkill
- resolve-conflictsSkill
- resolve-issue-commentsSkill
- resolve-pr-commentsSkill
- resolve-review-findingsSkill
- review-prSkill
- reviewSkill
- run-test-planSkill
- setup-worktreeSkill
- update-contextSkill
- await-pr-checksSkill
- commitSkill
- create-agentSkill
- create-issueSkill
- create-mcpSkill
- create-planSkill
- create-prSkill
- create-ruleSkill
- create-skillSkill
- create-test-planSkill
- implement-planSkill
- orchestrate-create-planSkill
- orchestrate-create-test-planSkill
- orchestrate-implement-planSkill
- orchestrate-resolve-pr-commentsSkill
- orchestrate-reviewSkill
- publish-releaseSkill
- resolve-conflictsSkill
- resolve-issue-commentsSkill
- resolve-pr-commentsSkill
- review-prSkill
- reviewSkill
- run-test-planSkill
- setup-worktreeSkill
- The first 60 shown.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
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.

