ardot-mcp
Skill4 starsChanged 7 days agoMITView on GitHub
This skill should be used when the user invokes /ardot-mcp or asks to package Ardot MCP through UXC, set up ardot-mcp-cli, complete Ardot OAuth for UXC, run sanitized Ardot readiness, open or inspect Ardot design files via a fixed CLI facade, or keep Grok/Claude native Ardot MCP as a compatibility path. Own Ardot endpoint identity, OAuth boundary, safe output, and task acceptance. Use uxc-facade for the generic packaging method. Do not use for one-off Ardot calls already solved by the native host MCP tools.
In this repository43 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- llms.txtllms.txt
- agent-symlink-initSkill
- code-inspector-initSkill
- commitSkill
- copywritingSkill
- css-tailwind-stylingSkill
- lingui-best-practicesSkill
- lingui-workflowSkill
- repo-skill-creatorSkill
- weekly-reportSkill
- agent-dispatchSkill
- agentic-review-handoffSkill
- agent-practice-auditSkill
- agent-symlink-initSkill
- architecture-hardening-loopSkill
- branch-creatorSkill
- chrome-dev-mcpSkill
- code-inspector-initSkill
- code-plugin-architectureSkill
- commit-pushSkill
- commitSkill
- decision-first-technical-writingSkill
- design-systemSkill
- desktop-agent-activitySkill
- dev-environment-roi-auditSkill
- disk-tidySkill
- figma-mcpSkill
- file-tidySkill
- goal-gateSkill
- grill-autopilotSkill
- kimi-computer-useSkill
- lingui-workflowSkill
- local-web-surfaceSkill
- open-code-review-loopSkill
- prompt-studioSkill
- review-prompt-composerSkill
- task-completion-loopSkill
- uxc-facadeSkill
- web-performance-auditSkill
- weekly-reportSkill
- workflow-gateSkill
- worktree-tidySkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 3 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 21
Treat /ardot-mcp without a task as readiness-only.
- Distinctive: line 27
If readiness reports missing OAuth or auth failure, run the OAuth handoff in references/oauth-and-binding.md, then retry readiness once.
- Very specific: line 28
If the managed CLI is absent, install and link only through this skill's scripts.
- Very specific: line 29
Use host-native Ardot MCP only when the user explicitly requests native compatibility or approves it after a blocker.
- Distinctive: line 56
Run only after the user authorizes local install and OAuth:
- Distinctive: line 67
Use this skill for:
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Ardot MCP
Use the managed ardot-mcp-cli facade for Ardot design-file work from a deterministic JSON CLI. This skill owns the Ardot endpoint, OAuth credential/binding, sanitized readiness, and task acceptance. Keep host-native Ardot MCP (for example Grok ardot-remote) as an explicit compatibility path, not the default shared CLI path.
This skill is the binary owner for pinned UXC 0.22.0. The UXC daemon under $HOME/.uxc is shared per user; do not let another skill silently install a different UXC version on the same machine. Use uxc-facade for the generic packaging method.
Keep the user contract simple
Let the user remember /ardot-mcp. Handle install, OAuth handoff, link repair, and readiness internally. With no additional task, return only ARDOT_MCP_READY after sanitized readiness. On failure, report one failing layer and one required action.
Execute the slash entry
Treat /ardot-mcp without a task as readiness-only. Treat the same invocation with a design task as readiness-plus-task.
Invoke every helper with zsh (for example zsh scripts/uxc-readiness.zsh). Under bash, set -u and ${0:A:h} abort with A: unbound variable.
- For readiness-only, run
zsh scripts/uxc-readiness.zshfrom this skill. It calls sharedsearch_style_guide, discards catalog candidates, and prints only bounded status fields. - For readiness-plus-task, run
zsh scripts/uxc-readiness.zsh --private-resultonce, keep the JSON private, and reuse that current-turn proof before design operations. - If readiness reports missing OAuth or auth failure, run the OAuth handoff in references/oauth-and-binding.md, then retry readiness once.
- If the managed CLI is absent, install and link only through this skill's scripts. Do not fall back to an unowned
uxconPATH. - Use host-native Ardot MCP only when the user explicitly requests
nativecompatibility or approves it after a blocker. Mark that modeNATIVE_COMPAT. - Claim a design task
VERIFIEDonly after the requested Ardot operation completes. Readiness alone is transport proof.
Preserve the runtime invariant
managed ardot-mcp-cli
-> pinned owned UXC 0.22.0
-> UXC OAuth credential + binding for ardot.tencent.com/mcp
-> Ardot MCP HTTP endpointFixed values:
| Item | Value |
| Endpoint | https://ardot.tencent.com/mcp |
| Link name | ardot-mcp-cli |
| Credential / binding id | ardot-mcp |
| Default scope | mcp:use |
| Default install/link dir | ${XDG_DATA_HOME:-$HOME/.local/share}/ardot-mcp |
| Binary owner | ardot-mcp |
HTTP MCP has no expensive stdio child, so this skill does not set --daemon-exclusive. Session reuse metadata may still appear; do not invent exclusivity keys.
Setup sequence
Run only after the user authorizes local install and OAuth:
zsh scripts/install-uxc.zshzsh scripts/setup-uxc-link.zsh- OAuth start/complete from references/oauth-and-binding.md
zsh scripts/uxc-readiness.zshtwice for install acceptance
zsh scripts/install-uxc.zsh --manifest prints pins without downloading.
Route by task semantics
Use this skill for:
- Installing or repairing the Ardot UXC facade and OAuth binding.
- Shared CLI discovery and sanitized readiness across agents.
- Design tasks that should go through
ardot-mcp-cliinstead of ad-hoc native tools.
For one-off inspection already connected in the current host's native Ardot MCP, stay on the native tools unless the user asked for the shared CLI facade.
Typical task operations after readiness:
ardot-mcp-cli -h
ardot-mcp-cli open_design fileUrl='https://ardot.tencent.com/file/<id>'
ardot-mcp-cli fetch_editor_state fileUrl='https://ardot.tencent.com/file/<id>' includeSchema=false includeGeneralEditInstructions=false
ardot-mcp-cli batch_read fileUrl='https://ardot.tencent.com/file/<id>'Confirm parameter names with ardot-mcp-cli <operation> -h. Never print OAuth tokens, raw readiness catalogs, or unrelated design payloads unless the user asked for that data.
Report acceptance
| Status | Proven by |
FACADE_READY | Owned binary + link contract gates passed |
TRANSPORT_READY | Sanitized search_style_guide through ardot-mcp-cli |
TASK_ACCEPTANCE | Requested Ardot operation completed |
NATIVE_COMPAT | Host-native Ardot MCP call, only when retained |
Mark unproved layers UNVERIFIED and name one next action. Read references/uxc-facade.md for ownership, pins, and evidence boundaries.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
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.

