Skip to main content
skill CX Codex

saving-workspace-context — Claude Code skill: project memory

The skill writes research, decisions and templates into your workspace files, so the next session starts with context instead of a blank slate.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add spencerpauly/awesome-cursor-skills --skill saving-workspace-context -a claude-code
Open on GitHub

What saving-workspace-context does

The agent acts as institutional memory for a project. At the start of a conversation it reads context/, a product context file such as .agents/product-marketing-context.md or PROJECT.md, domain folders (companies/, docs/, research/) and any templates that apply. No context/ directory yet? It creates one the first time there is something worth saving.

During the conversation it sorts findings by type: product details and positioning go to the project context file, research on a company or topic to context/{topic-slug}.md, strategy decisions to the same place with dated entries, reusable boilerplate to templates/. A multi-step workflow that will come up again gets proposed as a new skill in .cursor/skills/ or .agents/skills/, and a persistent constraint as a rule in .cursor/rules/. Small saves happen without asking — the agent just tells you what it wrote. Creating a skill or rule is different: it asks first, because those affect every future conversation. Existing files get appended to with dated entries rather than overwritten.

All of this lives inside the project workspace, so the agent needs write access there or there is nowhere to save. One more thing worth knowing: the skill decides on its own what counts as worth keeping. Leave it running for a while and context/ fills up with entries you may never read again.

What it helps with

  • rebuild project context in a fresh session without re-explaining everything
  • keep company research so it doesn't have to be redone
  • log a strategy decision with a date and find it later
  • move reusable boilerplate into templates/
  • turn a repeated multi-step workflow into its own skill

Example prompts for the agent

  • check what's already in context/ and pick up the landing page work
  • save to the project context that we switched the target audience to developers
  • turn our call prep process into a new skill

How to install saving-workspace-context

1Skills CLI (Vercel)

npx skills add spencerpauly/awesome-cursor-skills --skill saving-workspace-context -a claude-code

2Manually: put the files in place

~/.claude/skills/saving-workspace-context/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/saving-workspace-context/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about saving-workspace-context

How do I install saving-workspace-context in Claude?

Run npx skills add spencerpauly/awesome-cursor-skills --skill saving-workspace-context -a claude-code. The skill file then lands at ~/.claude/skills/saving-workspace-context/SKILL.md.

What does saving-workspace-context do?

It watches the conversation and writes useful material into project files: product context, research, decisions, templates. On the next session the agent reads those files, so you don't have to explain the project again.

Does the skill ask before saving anything?

Not for small context saves — it writes them and mentions what it saved. It does ask before creating a new skill or rule, since those change how the agent behaves in every future conversation.

Source text

SKILL.md · CC0-1.0
---
name: saving-workspace-context
description: Automatically persist useful context — research, decisions, learnings, templates — to workspace files so knowledge survives across conversations.
user-invocable: false
---

# Saving Workspace Context

You are an agent that builds institutional memory. As you work, watch for information that should outlast this conversation and save it to the workspace so future sessions start smarter.

## At the Start of Every Conversation

Load existing context before doing anything else:

1. Check for a `context/` directory — read any files relevant to the current task
2. Check for a product/project context file (e.g. `.agents/product-marketing-context.md`, `PROJECT.md`, or similar) for positioning, goals, and constraints
3. Check for any domain-specific directories the project uses (e.g. `companies/`, `docs/`, `research/`)
4. Check for templates or reusable assets that might apply

If the project doesn't have a `context/` directory yet, that's fine — create one when you first have something worth saving.

## During a Conversation

Watch for information that should be persisted. Save it as soon as you recognize it — don't wait until the end.

| Signal | Where to Save |
|---|---|
| Product details, positioning, ICP changes | Project context file (e.g. `.agents/product-marketing-context.md`) |
| Research on a company, person, or topic | `context/{topic-slug}.md` or a domain-specific directory |
| Strategy decisions or learnings | `context/{topic}.md` with dated entries |
| Reusable templates or boilerplate | `templates/` or a project-appropriate location |
| A repeatable multi-step workflow | New skill in `.cursor/skills/` or `.agents/skills/` |
| A persistent constraint or convention | New rule in `.cursor/rules/` |

### How to Save

- **Don't ask permission** for small context saves — just do it and mention what you saved
- **Do ask permission** before creating new skills or rules (they affect all future conversations)
- **Append, don't overwrite** when adding to existing context files — use dated entries
- **Use clear file names** — future you (or a future agent) needs to find this by scanning a directory listing

## At the End of a Conversation

Before finishing, ask yourself:

- Did I learn anything about this project that isn't captured in workspace files?
- Did I do research that would be painful to redo?
- Did I discover a pattern that should become a skill or rule?
- Did I create content that could be templated for reuse?

If yes to any, save it before the conversation ends.

## File Formats

### Context Files (`context/{slug}.md`)

```markdown
# {Topic}

## {Date} — {Brief title}

{What was learned, decided, or discovered}

## {Earlier date} — {Earlier entry}

{Previous context}
```

Keep entries reverse-chronological (newest first). Date your entries so they age gracefully.

### Project Context File

A single file capturing the current state of the project's identity:

```markdown
# {Project Name} — Context

- **What it is:** {one line}
- **Who it's for:** {target audience}
- **Key differentiator:** {why this vs alternatives}
- **Current stage:** {pre-launch / beta / growth / etc.}
- **Current goals:** {what matters right now}

## Positioning

{How we talk about the product}

## Constraints

{Things to always keep in mind}
```

## When to Create a New Skill

Create a skill when you find yourself doing the same multi-step workflow more than once:

- Researching a topic (check multiple sources, synthesize, save findings)
- Preparing for a meeting or call (pull context, recent history, prep talking points)
- Running a campaign or process (select targets, personalize, track progress)

## When to Create a New Rule

Create a rule when a persistent constraint should apply across all conversations:

- Voice/tone guidelines that get refined through feedback
- Naming conventions or file organization patterns
- Domain-specific constraints ("never mention X", "always check Y first")

## Rules

- Be proactive — save context without being asked, but mention what you saved
- Keep files scannable — future agents will skim directory listings to find context
- Don't save trivial information — if it's easily re-derived, skip it
- Date everything that accumulates over time
- Check for existing files before creating new ones to avoid duplicates

© spencerpauly · CC0-1.0 · spencerpauly/awesome-cursor-skills

Similar skills

All skills for Agent setup: memory, context, orchestration →

From the same set59

parallel-code-review Runs four parallel read-only subagents to inspect a diff across security, speed, correctness, and readability. Built for vetting large or risky pull requests. CX Codex +1 530installs reviewing-code A skill that walks your code through a checklist: correctness, maintainability, performance, types and tests. Useful when a change needs review before merge. CX Codex +1 516installs writing-tests Inspects existing code, detects your test runner, and drafts unit and integration tests with mocks and edge cases. CX Codex +1 499installs api-smoke-testing Finds API endpoints across your codebase, spins up the local dev server, and tests every route with curl. Spots unmounted 404 routes and 500 server crashes before release. CX Codex +1 495installs codebase-onboarding Fires five explore subagents in parallel to cover architecture, data models, APIs, auth and deployment, then merges their findings into one onboarding doc. CX Codex +1 476installs auditing-performance A skill that audits app performance: bundle size, rendering, database queries and Core Web Vitals. Useful when pages load slowly or the UI stutters. CX Codex +1 476installs suggesting-cursor-hooks Watches for repetitive check requests and suggests automating them with Cursor hooks. Useful when you find yourself repeatedly asking the agent to run linters or tests. CX Codex +1 475installs building-skills-from-patterns Turns repeated multi-step workflows into reusable SKILL.md files. It triggers when an agent re-derives the same routine or needs the same user correction three times. CX Codex +1 466installs

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor