Card

drwn card manages mind cards from authoring through project consumption.

Author and publish:

drwn card new @your-handle/backend --no-git
drwn card new backend --scope @your-handle --from-project .
drwn card publish @your-handle/backend
drwn card catalog publish @your-handle/[email protected] --catalog @your-handle --mode direct
drwn card show @your-handle/[email protected]
drwn card validate @your-handle/[email protected]
drwn card diff @your-handle/[email protected] @your-handle/[email protected]
drwn card deprecate @your-handle/[email protected]

Edit sources:

drwn card source list
drwn card source show @your-handle/backend
drwn card source show @your-handle/backend --json
drwn card source doctor
drwn card source doctor @your-handle/backend
drwn card source add-skill @your-handle/backend reviewer
drwn card source add-skill @your-handle/backend reviewer --from ./skills/reviewer --force
drwn card source add-skill @your-handle/backend reviewer --replace
drwn card source remove-skill @your-handle/backend reviewer
drwn card source remove-skill @your-handle/backend reviewer --keep-files
drwn card source set @your-handle/backend --description "Backend review harness" --version 0.1.0
drwn card source set @your-handle/backend --harness-min-version 0.4.0 --license MIT
drwn card source set @your-handle/backend --stability stable --last-validated-with 0.1.0 --test-status-badge https://example.com/status.svg
drwn card source add-mcp @your-handle/backend context7
drwn card source add-mcp @your-handle/backend context7 --from ./context7.json
drwn card source add-mcp @your-handle/backend context7 --replace
drwn card source remove-mcp @your-handle/backend context7
drwn card source remove-mcp @your-handle/backend context7 --keep-files

Mind content (beliefs, persona, memory, hooks):

# Beliefs — factual assertions composed into the system prompt
drwn card source add-belief @your-handle/mind engineering --visibility public
drwn card source remove-belief @your-handle/mind engineering
drwn card source remove-belief @your-handle/mind engineering --keep-files

# Persona — behavioral and stylistic identity
drwn card source add-persona @your-handle/mind voice --visibility internal
drwn card source remove-persona @your-handle/mind voice
drwn card source remove-persona @your-handle/mind voice --keep-files

# Memory — layered knowledge (l4 = short-term, l5 = mid-term, l6 = long-term)
drwn card source add-memory @your-handle/mind context --layer l4 --visibility private --format md
drwn card source add-memory @your-handle/mind raw --layer l6 --visibility private --format jsonl
drwn card source remove-memory @your-handle/mind context --layer l4
drwn card source remove-memory @your-handle/mind context --layer l4 --keep-files

# Hook policies — runtime tool intercept modules
drwn card source add-hook @your-handle/backend audit-tool-calls
drwn card source remove-hook @your-handle/backend audit-tool-calls
drwn card source remove-hook @your-handle/backend audit-tool-calls --keep-files

All add-* and remove-* source commands accept --dry-run and --json.

add-belief, add-persona, and add-memory require --visibility (private, internal, or public). add-memory additionally requires --layer (l4, l5, or l6) and accepts --format (md, jsonl, or mixed; default md).

remove-* commands delete the bundled directory by default. Pass --keep-files to remove only the manifest entry while keeping files on disk.

Consume cards in a project:

drwn apply @your-handle/backend@^1.0.0
drwn card apply @your-handle/backend@^1.0.0 --write
drwn card add @your-handle/backend@^1.0.0
drwn card add @your-handle/backend@^1.0.0 --allow-untrusted-source
drwn card pin @your-handle/[email protected]
drwn card remove @your-handle/backend
drwn card detach
drwn card update
drwn card outdated
drwn card status --explain

Every mutating consumer command accepts --write to chain into drwn write after the lock mutation succeeds:

drwn card add @your-handle/backend@^1.0.0 --write
drwn card pin @your-handle/[email protected] --write
drwn card remove @your-handle/backend --write
drwn card update --write
drwn card detach --write

drwn update is a top-level alias for drwn card update.

If a consumed card declares optional MCP servers, --write output reports whether each one is active, skipped, or shadowed by a different active definition. Skipped optional MCPs are not materialized until the project opts in:

drwn add mcp <server-name>
drwn write --dry-run

List local cards:

drwn card list
drwn card list --json

Check for outdated locks in CI and pre-fetch origins:

drwn card outdated --check
drwn card outdated --fetch
drwn card outdated --check --json

--check exits non-zero when any locked card has a newer version available locally, which makes it suitable as a CI gate. --fetch runs git fetch against each card's origin before computing the diff so the check uses up-to-date tag listings.

Remote and catalog flows use Git refs. drwn delegates authentication to Git.

Clone and fetch cards from Git remotes:

# Clone a card from a Git ref into the local store
drwn card clone git+https://github.com/team/backend.git#v1.0.0
drwn card clone github:team/backend@^1.0.0
drwn card clone git+https://github.com/external/card.git#v1.0.0 --allow-untrusted-source

# Fetch updates for a locally cloned card from its remote
drwn card fetch @team/backend
drwn card fetch @team/backend --remote upstream

Manage Git remotes for a local card repo:

# Add a new remote
drwn card remote add @team/backend https://github.com/team/backend.git

# Add or update the origin URL (idempotent)
drwn card remote set @team/backend https://github.com/team/backend.git

# Use a non-default remote name
drwn card remote set @team/backend https://internal.example.com/backend.git --name upstream

# Remove a remote
drwn card remote remove @team/backend

# List configured remotes
drwn card remote list @team/backend
drwn card remote list @team/backend --json

Push a card to its configured remote:

# Push main branch and all version tags
drwn card push @team/backend

# Push to a non-default remote
drwn card push @team/backend --remote upstream

# Declare the remote visibility to control the push gate
drwn card push @team/backend --remote-visibility private
drwn card push @team/backend --remote-visibility public --unsafe-push-public

drwn card push evaluates a visibility gate before pushing. If the card contains content with a stricter visibility than the target remote (e.g., private mind content pushed to a public remote), the push is blocked. Use --remote-visibility to declare the remote's visibility when drwn cannot auto-detect it, and --unsafe-push-public to override the block when you intentionally want to publish restricted content.

Publish to a card catalog after the card has been pushed to an installable Git remote:

drwn card remote add @team/backend <card-git-url>
drwn card push @team/backend
drwn library catalog add <catalog-git-url>
drwn card catalog publish @team/[email protected] --catalog @team --mode direct --tag backend --json
drwn search card backend --scope @team

drwn card catalog publish accepts a store card ref or Git-origin card ref. --catalog can be a registered scope such as @team, a catalog Git URL, or a local catalog checkout path. --mode local updates catalog.json only; it does not commit or push. --mode direct requires a clean catalog worktree, commits the catalog.json change, pushes the current branch, and refreshes a registered catalog cache when possible.

Use --dry-run --json to validate the card ref, entry URL, catalog schema, and duplicate-entry behavior before writing. Existing entries require --replace unless the generated payload is already identical.

Hook Consent

When a locked card declares hook policies, drwn write will not materialize hooks until consent is explicitly recorded. drwn doctor surfaces this as hookIssues.

Review and grant consent:

drwn card trust @your-handle/backend --hooks

By default, consent covers ^<locked-version>. Override with --range:

drwn card trust @your-handle/backend --hooks --range "^1.0.0"

Revoke consent:

drwn card untrust @your-handle/backend --hooks

Preview what hooks a card declares without granting consent:

drwn card audit

drwn card audit is a v1.1 placeholder — it makes the planned command discoverable but does not yet produce a diff output.

Typical Source Authoring

The canonical authoring sequence from empty source to published card:

drwn card new @your-handle/backend --no-git
drwn card source add-skill @your-handle/backend reviewer
drwn card source add-mcp @your-handle/backend context7
drwn card source set @your-handle/backend \
  --description "Backend review harness" \
  --version 0.1.0 \
  --stability stable \
  --last-validated-with 0.1.0 \
  --test-status-badge https://example.com/status.svg
drwn card source doctor @your-handle/backend
drwn card publish @your-handle/backend

The quality fields (--stability, --last-validated-with, --test-status-badge) surface in drwn card show so consumers can see the maturity signal before applying a card.