Using Paqad with Cursor

In Cursor, Paqad is advisory: onboarding writes a rule file that tells Cursor's AI how to load your project's rules and workflows, but nothing in Cursor makes the AI follow it. Paqad installs no hooks for Cursor, so no Paqad program runs before an edit or when the AI finishes.

What Paqad can enforce in Cursor

Nothing, during the session. Cursor is at the advisory level: the tool is given the rules but nothing stops it ignoring them. In practice this means:

• No edit is ever refused by Paqad.
• No verification runs when the AI finishes.
• Paqad's evidence files are only written if the AI chooses to run Paqad's commands itself.

Paqad's source records Cursor as advisory with hooks switched off for this adapter.

Files onboarding writes

File	What it is
.cursor/rules/paqad.mdc	The entry file, written as a Cursor project rule. It tells the AI to open .paqad/framework-path.txt and follow the installed framework, or to act as a normal assistant if Paqad is missing or off.
.cursor/mcp.json	The MCP servers Paqad suggests for your stack. MCP (Model Context Protocol) lets AI tools connect to extra information sources. This file names servers but holds no start commands.
.cursor/cache.json, .cursor/memory.json	Paqad bookkeeping. Not Cursor settings.

Onboarding does not create .cursorrules, does not write one rule file per topic, and does not copy skills or agents into .cursor/. Paqad's skills and workflows stay in the installed framework folder.

Check that Cursor applies the rule

The generated paqad.mdc contains plain instructions and no settings block at the top saying when it should apply. After onboarding, open Cursor's rules settings and check how the paqad rule is listed. If it is not applied to every chat, mention the rule in your message or change its type in Cursor. Be aware that paqad-ai refresh --providers and paqad-ai update can rewrite the file.

Getting evidence without hooks

Paqad's instructions tell the AI, on advisory tools, to record each stage of feature work itself by running Paqad's commands in the terminal, for example:

npx paqad-ai stage start planning
npx paqad-ai stage end planning --artifact <plan file>

These commands write to the ledger (Paqad's evidence files under .paqad/ledger/) exactly as they would on any other tool. Whether the AI runs them is up to the AI. To see what really happened, look in .paqad/ledger/paqad.stage-evidence/ after a change. An AI that only describes the stages in chat, without running the commands, leaves the ledger empty.

Two things do run regardless of the AI. Git post-commit and post-merge hooks link commits to their feature, and you can add Paqad's verification script to your CI pipeline. See Adapters overview.

Setting it up

1. Back up .cursor/mcp.json if you have one.
2. Run paqad-ai onboard --providers cursor, or tick Cursor when onboarding asks.
3. Check the rule type in Cursor, as described above.

If you need edits to be refused when the process is skipped, use Claude Code or Codex CLI.