Refresh

Skill joaquimscosta/arkhe-claude-plugins/plugins/roadmap/skills/refresh

Supercharge Claude Code with 109 specialized components — 22 agents, 32 commands, 55 skills across 13 modular plugins. Deep reasoning, autonomous dev loops, DDD architecture, design system enforcement, git automation, and more.

Install

npx -y skills add joaquimscosta/arkhe-claude-plugins --skill refresh

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

21 stars21 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Detect drift in .arkhe/roadmap/ context files and regenerate them from the current codebase state. Scaffolds context directory if missing. Use when user runs /roadmap:refresh, mentions "refresh context", "update context", "init roadmap", "stale context", or "scaffold roadmap".

SKILL.md

3.7 KB, 827 tokens by cl100k_base, as published. Nobody here has run it

Context Directory Refresh

Detect drift in .arkhe/roadmap/ context files and regenerate them using a hybrid format (condensed summary + references to authoritative docs).

Arguments

Parse from $ARGUMENTS:

Mode	Description
`init`	Scaffold `.arkhe/roadmap/` with all 3 files from scratch
`check`	Detect drift only — report staleness, don't write
`all`	Regenerate all 3 context files
`project`	Regenerate only project.md
`architecture`	Regenerate only architecture.md
`documents`	Regenerate only documents.md
(none)	Run `check`, then ask which files to refresh

Step 1: Run Drift Detection

Run the detection script:

python3 ${CLAUDE_PLUGIN_ROOT}/skills/refresh/scripts/detect_context_drift.py .

Parse the JSON output. Present a drift report table:

## Context Directory Status

| File | Exists | Staleness | Days | Commits Since | feat/fix |
|------|--------|-----------|------|---------------|----------|
| project.md | Yes/No | fresh/aging/stale/very_stale | N | N | N |
| architecture.md | Yes/No | ... | ... | ... | ... |
| documents.md | Yes/No | ... | ... | ... | ... |

New modules not in architecture.md: {list or "none"}

For check mode: Stop here. Present the report and suggest which files need attention.

Step 2: Context Discovery

For init, all, or single-file modes, run the shared context discovery protocol from CONTEXT_DISCOVERY.md. Gather:

From README.md: project purpose, target users
From CLAUDE.md: constraints, conventions, tech stack
From build files: tech stack detection via TECH_STACK_DETECTION.md
From docs/**/*.md: documentation inventory
From module globs: module inventory with file counts
From docs/adr/**/*.md: ADR titles and status
From arkhe/specs/*/spec.md: spec pipeline

Step 3: Generate Files

Generate each requested file using the hybrid format below. See WORKFLOW.md for detailed generation protocol per file.

Hybrid Format Template

# {Title}
_Auto-generated by /roadmap:refresh on {date}. Edit freely — regenerate with `/roadmap:refresh {file}`._

## Summary
{3-5 bullet points, <300 tokens — the essential facts}

## Details
{Structured data: tables, lists}

## References
{Pointers to authoritative docs}
- `{doc_path}` — {one-line description}

Generation Sources

File	Primary Sources
`project.md`	README.md, CLAUDE.md, gap analyses, existing project.md
`architecture.md`	Build files, module globs, ADRs, CLAUDE.md architecture section
`documents.md`	Phase 5 doc scan results, spec files, gap analyses, ADRs

Step 4: Confirm and Write

Before writing each file:

Show current content (if exists) or "Does not exist"
Show proposed new content
Ask: "Write this to {context_dir}/{filename}?"

Create the .arkhe/roadmap/ directory if it doesn't exist.

Never overwrite without confirmation.

References

WORKFLOW.md — Detailed generation protocol per file
EXAMPLES.md — Usage examples
TROUBLESHOOTING.md — Common issues

Gives 0 of the 12 instructions most project setup skills give in 827 tokens

Counted across 999 of the 1,637 authors here whose files we hold, read 2026-08-06

ask one question at a timein 29 of 999, across 28 files
detect the package manager from lockfilesin 28 of 999, across 9 files
present findings to the userin 25 of 999, across 4 files
explore current repo statein 24 of 999, across 3 files
update the agent skills block in place if it existsin 24 of 999, across 3 files
install husky lint-staged and prettierin 23 of 999, across 4 files
create the lintstagedrc filein 22 of 999, across 3 files
commit all changed filesin 22 of 999, across 3 files
run lint-staged to verify it worksin 22 of 999, across 3 files
initialize huskyin 21 of 999, across 2 files
create the husky pre-commit filein 21 of 999, across 2 files
create a prettierrc file if missingin 21 of 999, across 2 files

Said here and by no other author read

Run the drift detection script
Present a drift report table
Gather context from sources
Generate requested files using hybrid format
Show current file content
Show proposed new content

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.