Skip to content
Agentmart
Submit a tool

compile-intent

Build a structured intent spec through guided conversation, one decision at a time with a recommended answer for each.

Skill by pathmodeio

IndexedMIT★ 3
Write a review

Chalk Labs

Need help deploying compile-intent?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/pathmodeio-compile-intent

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Build a structured intent spec through guided conversation, one decision at a time with a recommended answer for each. Use when the user wants to define what to build, sharpen a vague idea into a testable spec, or capture product intent before writing code. Writes intent.md in the project root as a local draft; a Pathmode workspace sees it after adoption.

Use cases: writing

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md and code. Checked Oct 26, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/compile-intent

What's inside

Skill name
compile-intent
Description
Build a structured intent spec through guided conversation, one decision at a time with a recommended answer for each. Use when the user wants to define what to build, sharpen a vague idea into a testable spec, or capture product intent before writing code. Writes intent.md in the project root as a local draft; a Pathmode workspace sees it after adoption.
License
MIT
Instructions
71 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

<what-to-do>

Invoke the `compile-intent` MCP prompt from the @pathmode/mcp-server. This starts a guided conversation that turns a vague problem into a structured intent spec.

For each question you ask, propose your best-guess answer based on the conversation so far. Don't make the user generate from a blank page.

Before the spec is finished, gather implementation context. You are already sitting in the repo, so read it: which files and modules this change lands in, what already exists there, what it would touch, and how it could be verified. Pass it to `intent_save` as `implementationContext` — in both modes. A brand-new spec has no intent id yet, so there is nothing to address a separate call to; `intent_save` carries the context through as part of the save. Use `record_implementation_context` only to update the context on an intent that already exists. It is free text, sub-headings welcome, and it is advisory — it never changes the readiness verdict. Its job is to stop the implementing agent from rediscovering the codebase from scratch, and to catch the case where the thing being specified half-exists already.

Save the proposal before pretending its open choices are requirements. Use the typed `productChoices` input on `intent_save` for consequential unanswered behavior; supply only proposals, never state, human actors, timestamps, or source-settlement flags. Existing choice IDs are stable; do not replace a question by recycling its ID. Word each contingent outcome claim as an observable result (a state, a number, a visible element) before the first save, because a saved choice cannot be edited: accepting a choice adds its outcomes to the spec, and the preflight flags any it cannot read as measurable, since such an outcome counts against the outcomes check. If one is flagged, say so when you explain the choice; the answer can then carry a measurable outcome instead (`answer_product_choice` action `change` without a workspace, the reviewer's change with one). The server preserves existing choices. A failed capability check means this installed path is unavailable, not that the question was recorded.

For a new choice-bearing proposal, call `intent_save(localDraft: true)` to write `intent.md` locally, then run `npx @pathmode/mcp-server@latest adopt` in the repository and relay its link and code for browser approval. Adoption preserves the open choices and establishes repository ownership. Ordinary v1 creates remain cloud-owned and reject choice proposals. Do not detach an already connected file with `localDraft`; add proposals through its normal connected save once it has repository authority. After successful adoption, call `attach_original_request`, copying the original request verbatim, and invite the reviewer to the adopted proposal. Stop while its choices remain unresolved or its exact revision lacks human authorization. Answers create corrections for the repository agent to apply; application and authorization are separate steps.

</what-to-do>

<supporting-info>

## Why a conversation, not a template

The compile-intent prompt is interrogative on purpose. It pushes back on vague language, challenges unmeasurable outcomes, and forces concrete constraints before moving on. Specs written in one shot tend to be wishful; specs that survive grilling are agent-ready.

## Output shape

`intent.md` at the project root, abridged (`intent_save` writes it; a section appears only when it has content):

```markdown
---
id: "intent_..."
version: 1
status: "draft"
---
# [Title]

## Objective
[What needs to change and for whom]

## Implementation Context
[What the repo actually looks like where this lands — read from the working tree, not guessed]

## Outcomes
- [ ] [Observable state change, testable in under 5 minutes]

## Constraints
- [Hard limit — what must never happen]

## Edge Cases
- **[Scenario]**: [Expected behavior]

## Verification
**[Check kind]**:
- [ ] [A check concrete enough to run without asking the author]
```

A new spec is saved as a local draft and stays on this machine, even with `PATHMODE_API_KEY` set. Adoption (`npx @pathmode/mcp-server@latest adopt`, approved in the browser) brings it into the Pathmode workspace, where team members and their agents can see it. Once `intent.md` is connected, each save also syncs to the workspace.

## Evidence in a file that lives in git

`intent.md` is committed, so everyone who clones the repo can read it forever and deleting a line later does not remove it from history. Under the IntentSpec repo-safe profile the file must not carry raw customer quotes, names or contact details of individuals, support transcripts, confidential internal metrics, or credentials. Cite evidence as a summary or a reference instead: a ticket id, a dashboard name, a one-line paraphrase. Full evidence with quotes and attribution belongs in the private system of record (a connected Pathmode workspace), and the repo file points back at it. A claim with no evidence yet is labelled an assumption, not dressed as a fact.

## Downstream skills

After compiling, these skills consume the spec:

- `grill-intent` — re-enter the conversation to find weaknesses before code gets written
- `review-against-intent` — check code changes against outcomes and constraints
- `handoff-intent` — capture decisions and discoveries back to the spec at end of session

</supporting-info>

If adoption reports `AUTHORIZATION_PENDING`, preserve the draft and rerun the same command after the person approves.

From pathmodeio/claude-plugin, MIT.

Signals

Stars
3
Forks
0
Open issues
0
Last commit (this folder)
Oct 9, 2026 (21 days ago)
Last release
Sep 7, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from pathmodeio

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get compile-intent running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying compile-intent?

CallGet help
Cite this page

compile-intent skill. Agent Mart. https://agentmart.dev/skills/pathmodeio-compile-intent (last updated 2026-10-24)