From accelerator
Initializes repo with directories like plans/research/decisions/prs and gitignore entries for Accelerator skills. Safe to run repeatedly, reports created vs existing.
npx claudepluginhub atomicinnovation/accelerator --plugin acceleratorThis skill is limited to using the following tools:
You prepare a consumer repository with all directories and gitignore entries
Detects single-project or monorepo structure, creates .groundwork.yml config for monorepos, updates .gitignore, and persists project selection for Groundwork.
Scaffolds AgenticDev directory structure (docs/specs, docs/changes) via bash script for spec-driven development workflow. Use when starting new projects or adding to existing ones without docs.
Initializes a git repository in the current directory, optionally adds mandatory commit instructions for agents to AGENTS.md, and generates a .gitignore for flavors like Node.js, Python, or custom patterns.
Share bugs, ideas, or general feedback.
You prepare a consumer repository with all directories and gitignore entries that Accelerator skills expect. This is safe to run repeatedly — it reports what was created versus what already existed.
Resolve each output directory using the plugin's path configuration:
Plans directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh plans meta/plans
Research directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh research meta/research
Decisions directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh decisions meta/decisions
PRs directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh prs meta/prs
Validations directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh validations meta/validations
Review plans directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh review_plans meta/reviews/plans
Review PRs directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh review_prs meta/reviews/prs
Review work items directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh review_work meta/reviews/work
Templates directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh templates meta/templates
Work items directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh work meta/work
Notes directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh notes meta/notes
Design inventories directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh design_inventories meta/design-inventories
Design gaps directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh design_gaps meta/design-gaps
Tmp directory: !${CLAUDE_PLUGIN_ROOT}/scripts/config-read-path.sh tmp meta/tmp
Execute the following steps in order, tracking what was created versus what already existed so you can present a summary at the end.
For each of the directories resolved above, run:
mkdir -p {directory}
Then check whether .gitkeep already exists. If not, create it:
touch {directory}/.gitkeep
Track the status of each directory:
mkdir -p: createdTo determine this, check whether the directory exists before running mkdir -p
(use [ -d {directory} ]).
The tmp directory needs a .gitignore that ignores everything except itself.
This inner .gitignore is the sole mechanism for ignoring tmp contents — do
NOT add the tmp path to the consumer's root .gitignore (a root-level ignore
entry would prevent git from descending into the directory, which means the
inner .gitignore's exclusions would never take effect and the directory would
not survive fresh clones).
Check whether {tmp directory}/.gitignore already exists. If not, write it
using the Write tool with this exact content:
# Ignore everything in this directory except the directory itself
*
!.gitkeep
!.gitignore
Track whether it was created or already exists.
.claude/accelerator.local.md to root .gitignoreCheck whether a .gitignore file exists at the project root.
If no .gitignore exists, create one containing:
.claude/accelerator.local.md
If .gitignore already exists, read it and check whether
.claude/accelerator.local.md is already present. When checking, trim
whitespace from both the existing lines and the candidate entry before
comparison. If not present, append it to the end of the file (with a leading
newline if the file does not end with one).
Track whether it was added or already present.
Only after all steps above have completed successfully, present a summary:
Initialisation complete:
Directories:
✓ {plans directory} (created | already exists)
✓ {research directory} (created | already exists)
✓ {decisions directory} (created | already exists)
✓ {prs directory} (created | already exists)
✓ {validations directory} (created | already exists)
✓ {review plans directory} (created | already exists)
✓ {review prs directory} (created | already exists)
✓ {review work items directory} (created | already exists)
✓ {templates directory} (created | already exists)
✓ {work items directory} (created | already exists)
✓ {notes directory} (created | already exists)
✓ {design inventories directory} (created | already exists)
✓ {design gaps directory} (created | already exists)
✓ {tmp directory} (created | already exists)
Tmp directory:
✓ {tmp directory}/.gitignore (created | already exists)
✓ {tmp directory}/.gitkeep (created | already exists)
Gitignore entries:
✓ .claude/accelerator.local.md (added | already present)
Use the actual resolved paths in the output (not the variable names).
/accelerator:init bootstraps a fresh repository. To upgrade an existing repository after a plugin update, use /accelerator:migrate instead — it applies ordered, idempotent migrations to bring meta/ and .claude/accelerator*.md in line with the latest schema.