docs(automation): add structured Automation, Skills, and nix-claude-code surface#30
Open
JacobPEvans wants to merge 5 commits into
Open
docs(automation): add structured Automation, Skills, and nix-claude-code surface#30JacobPEvans wants to merge 5 commits into
JacobPEvans wants to merge 5 commits into
Conversation
|
Preview deployment for your docs. Learn more about Mintlify Previews.
|
There was a problem hiding this comment.
Code Review
This pull request introduces a comprehensive "Automation" section to the documentation, detailing the mechanics of the issue-to-mergeable PR pipeline. It adds several new pages covering ai-workflows, claude-code-routines, codeql-resolution, shape-issues, and the /ship and /finalize-pr skills, while updating navigation and cross-references across existing files. Feedback from the review highlights a timezone inconsistency in the routine schedule and a count mismatch in the automation overview heading.
JacobPEvans
added a commit
that referenced
this pull request
May 24, 2026
…g parts Addresses gemini-code-assist review on PR #30: - claude-code-routines.mdx: README in source repo lists all routines in CT, not UTC. Updated table header to "Schedule (CT)" and corrected every row to match the source-of-truth times from the README. - overview.mdx: heading said "seven moving parts" but the CardGroup has six cards. Changed to "six moving parts". Assisted-by: Claude <noreply@anthropic.com>
Adds a new top-level Automation nav group with seven pages covering the end-to-end issue->mergeable-PR pipeline that spans three repos: - automation/overview — the three-layer pipeline (cloud GHA, cron routines, local /ship) with a canonical end-to-end diagram - automation/issue-to-pr-pipeline — the eight-step cloud flow on this repo, with which trigger fires which caller and what each does - automation/ai-workflows — catalog of the 16 reusable workflows in JacobPEvans/ai-workflows: event-triggered, scheduled, and what's deprecated or disabled - automation/claude-code-routines — the six cron-scheduled cloud routines: morning briefing, sentinel, custodian, issue solver, daily polish, weekly scorecard; the deploy-routines GHA workaround - automation/ship-and-finalize — /ship + /finalize-pr lifecycle with the canonical five-phase diagram and the known reliability gaps being closed in a follow-up against claude-code-plugins - automation/codeql-resolution — why CodeQL is fixed separately from CI checks, and how /resolve-codeql plugs into /finalize-pr Phase 2 - automation/shape-issues — the /shape-issues local skill and how it feeds well-scoped issues into the resolver pipeline Cross-links from existing pages: - architecture/ai-pipeline.mdx -> /automation/overview (mechanics) - tools/automation.mdx -> /automation/overview + /automation/ship-and-finalize - ai-development/claude-code-plugins.mdx -> /automation/ship-and-finalize + replaces direct GitHub links with /automation/* internal links for claude-code-routines and ai-workflows All new diagrams use the canonical hand-drawn Reef Green theme directive byte-for-byte; grep validates one unique directive across the site. Word counts within tier caps (tier 1 <=450, tier 2 <=900). All internal links validated by mint broken-links. Assisted-by: Claude <noreply@anthropic.com>
…g parts Addresses gemini-code-assist review on PR #30: - claude-code-routines.mdx: README in source repo lists all routines in CT, not UTC. Updated table header to "Schedule (CT)" and corrected every row to match the source-of-truth times from the README. - overview.mdx: heading said "seven moving parts" but the CardGroup has six cards. Changed to "six moving parts". Assisted-by: Claude <noreply@anthropic.com>
…I key ai-workflows now authenticates directly against the Anthropic API using ANTHROPIC_API_KEY, drawing from the Claude Max plan's $200/month automation credits. OpenRouter remains supported as a fallback provider but is no longer the default. - ai-workflows.mdx: Authentication section rewritten; secret list swapped (ANTHROPIC_API_KEY replaces OPENROUTER_API_KEY + OPENROUTER_BASE_URL); fallback mention kept for completeness; "Where to go next" cards updated. - issue-to-pr-pipeline.mdx: secrets table updated; auth card description matches the new default. Assisted-by: Claude <noreply@anthropic.com>
c6845cb to
6826e93
Compare
… add nix-claude-code Restructure the Automation group for future growth: - automation/cloud-pipelines/ (NEW subdir) — ai-workflows lives here now - automation/scheduled-routines/ (NEW subdir) — claude-code-routines lives here - ai-development/skills/ (NEW subdir) — ship-and-finalize, codeql-resolution moved out of automation/ since they are local Claude skills, not pipeline plumbing - nix/nix-claude-code (NEW page) — declarative Claude Code via home-manager, the missing piece between the binary (nix-ai) and the install (this) - automation/shape-issues DELETED — /shape-issues is not actually used; Claude plan mode covers the same workflow Each new subdirectory carries its own overview.mdx as a landing page. The top-level automation/overview now points at three pillars (Cloud pipelines, Scheduled routines, Skills) instead of a flat seven-card grid. Also scrub the Claude Max plan / $200/month pricing line in ai-workflows.mdx (replaced with "included monthly automation credits on the Claude subscription that backs these workflows"), and codify the cost-and-tier reporting rule in security/scrubbed-values.mdx: - Never name a specific AI subscription tier in committed content - Cost figures are monthly with the per-instance × instances/month math visible - Schedules and rate limits are not costs and stay as-is Also fix two stale agentsmd/rules/* references that were broken by the prior rules-migration PR: - automation/cloud-pipelines/ai-workflows.mdx: ci-cd-policy.md link now points at /infrastructure/cicd/policy - automation/issue-to-pr-pipeline.mdx: git-signing.md link now points at /infrastructure/cicd/git-signing 27 internal links audited; all resolve. JSON valid. Mermaid theme directive remains the single canonical form across the site. Assisted-by: Claude <noreply@anthropic.com>
The section listed marketing tiers (Claude Max, ChatGPT Pro, Copilot Business) by name as a forbidden list, which made the public docs site the public memorial of a behavioral rule that belongs in agent context, not on a docs page. The underlying behavior — no specific pricing or tier names in committed content — stays as agent guidance. Assisted-by: Claude <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
cloud-pipelines/andscheduled-routines/instead of a flat list, so the surface can absorb future growth without becoming a wall of siblings.ai-development/skills/for local Claude Code slash-commands (/ship,/resolve-codeql). These were briefly underautomation/before review — they are agent tooling, not pipeline plumbing, so they live in AI Development now.nix/nix-claude-codedocumenting the declarative Claude Code home-manager modules indryvist/nix-claude-code, with explicit cross-links to the rest of the Nix stack (nix-darwin, nix-home, nix-ai, nix-devenv).security/scrubbed-values.mdx: no specific subscription-tier names, no$N/day, costs are monthly with per-instance × instances/month math.What's in the new nav
overview,issue-to-pr-pipeline,cloud-pipelines/{overview, ai-workflows},scheduled-routines/{overview, claude-code-routines}skills/{overview, ship-and-finalize, codeql-resolution}nix-claude-codeWhat's deleted
automation/shape-issues.mdx— the/shape-issuesskill is not actually in use; Claude plan mode covers the same workflow.Content scrubs
automation/cloud-pipelines/ai-workflows.mdx:Claude Max plan's $200/month automation credits→ vagueincluded monthly automation credits on the Claude subscription that backs these workflows.agentsmd/rules/*.mdlinks (broken when docs: migrate cross-repo standards from ai-assistant-instructions #32 migrated those rules) repointed at/infrastructure/cicd/policyand/infrastructure/cicd/git-signing.Verification
jq empty docs.json✓/automation/<old-name>references ✓$200,Max plan,Pro plan,$N/dayleaks (the policy text inscrubbed-values.mdxitself is the only intentional match) ✓mergeStateStatus: CLEAN✓Test plan
mint dev— confirm the Automation sidebar shows the two new sub-groups (Cloud pipelines, Scheduled routines), AI Development shows the new Skills sub-group, and Nix Ecosystem listsnix-claude-codebetweennix-devenvandpackage-placement./automation/overview— three pillars now (Cloud pipelines, Scheduled routines, Skills)./automation/cloud-pipelines/ai-workflowsno longer contains theClaude Max plan/$200/monthwording.mint broken-links.Refs: #32