[step-names] Align two source step names: MempPalace MCP Server casing and Cache→Save pattern

[github-actions]

Step Name Alignment Issues

Found two source-level step name inconsistencies that drift from previously established conventions. Both live in source .md files (so they propagate through every workflow that imports or hosts them).

Summary

• One glossary-capitalization mismatch: MCP server should be MCP Server per the project glossary.
• One pattern regression: two workflows still use the old Cache <X> step name even though the project standardized on Save <X> to cache in closed issue [step-names] Align "Cache (X)" step names with "Save (X) to cache" pattern #23941.

Issues Identified

1. [Medium Priority] Glossary capitalization: MCP server → MCP Server

Current step name:

• .github/workflows/shared/mcp/mempalace.md:46 — Start MemPalace MCP server

The rendered name also appears in the compiled workflow that consumes this shared import (e.g. .github/workflows/daily-fact.lock.yml).

Issue:
The project glossary defines the term as MCP Server (capital S). Every other related step name already follows this casing, for example:

• Generate Safe Outputs MCP Server Config
• Determine automatic lockdown mode for GitHub MCP Server
• Start MCP Scripts HTTP Server
• Start Safe Outputs MCP HTTP Server
• Start MCP Gateway / Stop MCP Gateway

Only the MemPalace shared step uses lowercase server, making it the odd one out.

Suggested improvement:

• Start MemPalace MCP server → Start MemPalace MCP Server

Glossary reference: ### MCP Server — defined with capital S.

---

2. [Medium Priority] Pattern regression: Cache <X> should be Save <X> to cache

Current step names:

• .github/workflows/layout-spec-maintainer.md:26 — Cache layout spec data
• .github/workflows/prompt-clustering-analysis.md:38 — Cache prompt clustering data

Both render unchanged into the corresponding .lock.yml files.

Issue:
Closed issue #23941 ("Align Cache (X) step names with Save (X) to cache pattern") standardized this verb pattern across the project. The cache-memory step family already follows it:

• Save cache-memory to cache (default)
• Save cache-memory to cache (focus-areas)
• Save cache-memory to cache (repo-audits)

The two workflows above are the only remaining holdouts using the older Cache <X> form.

Suggested improvement:

• Cache layout spec data → Save layout spec data to cache
• Cache prompt clustering data → Save prompt clustering data to cache

Related closed issue: #23941.

---

Agentic Task Description

To align these step names:

1. Update the source .md files (not the .lock.yml):
   • .github/workflows/shared/mcp/mempalace.md line 46 → MCP Server
   • .github/workflows/layout-spec-maintainer.md line 26 → Save layout spec data to cache
   • .github/workflows/prompt-clustering-analysis.md line 38 → Save prompt clustering data to cache
2. Recompile each affected workflow with gh aw compile <workflow>.md (and any workflow that imports shared/mcp/mempalace.md, e.g. daily-fact).
3. Verify the regenerated .lock.yml files show the new names.
4. No behavioral changes — these are display-only string edits, so smoke testing should be a quick compile + lint pass.

Related Files

• Source: .github/workflows/shared/mcp/mempalace.md
• Source: .github/workflows/layout-spec-maintainer.md
• Source: .github/workflows/prompt-clustering-analysis.md
• Compiled (auto-regenerated on recompile): .github/workflows/daily-fact.lock.yml, .github/workflows/layout-spec-maintainer.lock.yml, .github/workflows/prompt-clustering-analysis.lock.yml
• Project glossary: docs/src/content/docs/reference/glossary.md
• Prior pattern issue: [step-names] Align "Cache (X)" step names with "Save (X) to cache" pattern #23941

Priority

Medium Priority — neither finding is user-facing or behavioral, but both diverge from explicit conventions the project has already standardized, so fixing them keeps step names predictable in logs and the Actions UI.

---

AI generated by Step Name Alignment for daily maintenance

Generated by 📋 Step Name Alignment · opus47 10.1M · ◷

• expires on May 30, 2026, 5:24 AM UTC

[github-actions]

🍪 Issue Monster selected this for Copilot

I've identified this issue as a good candidate for automated resolution and requested assignment to the Copilot coding agent.

If assignment succeeds, the Copilot coding agent will analyze the issue and create a pull request with the fix.

Om nom nom! 🍪

🍪 Om nom nom by Issue Monster · haiku45 31.7K · ◷