setup-superconductor-project
Skill451 starsChanged 7 months agoMITView on GitHub
Configures a repository for super.engineering with repo-owned worktree setup and development-server scripts. Use when asked to "make this project Superconductor ready", "set up super.engineering", "add worktree setup", or "configure the dev server".
- Installs packages
We split this skill into the rules it gives an agent
13 rules. About 0 are common, 8 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 14
Run sc instructions project and sc project scripts --help before mutation.
- Distinctive: line 24
Treat live instructions and current help as authoritative.
- Very specific: line 28
Read repository instruction files, the existing .superconductor/config.json when present, root manifests, lockfiles, workspace …
- Very specific: line 33
Prefer an existing dev or start script for the default run target.
- Distinctive: line 37
Ask one focused question when the package manager, run script, or intended service is ambiguous.
- Distinctive: line 39
Do not invent teardown, stop, or cleanup commands.
- Very specific: line 43
Use repo scope so managed worktrees inherit the configuration.
- Distinctive: line 50
If no setup command is needed, do not add an empty one.
- Distinctive: line 60
Run the setup command once when it is safe and non-destructive.
- Very specific: line 61
Start the configured target through sc project scripts run <name> --json.
- Distinctive: line 63
When the service exposes HTTP and its port is known from evidence, request its IPv4 loopback URL and require a successful response.
- Distinctive: line 64
Stop the target after verification unless the user asked to leave it running, then verify it is idle.
- Distinctive: line 89
Leave the configuration uncommitted unless the user separately requests a commit.
≈ “Common” is an estimate: at least 1 in 5 of the 308,232 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
Set Up a Superconductor Project
Configure the current repository's .superconductor/config.json through the supported sc project scripts interface. Do not create a worktree, commit, push, or open a browser unless separately requested.
Step 1: Load Live Guidance
- Read
~/.pi/agent/skills/superconductor/SKILL.md. - Run
sc instructions projectandsc project scripts --helpbefore mutation. - Inspect current state:
sc project scripts --json
sc project scripts check --json
sc worktree status --json
git status --short --branchTreat live instructions and current help as authoritative.
Step 2: Derive Commands from Repository Evidence
Read repository instruction files, the existing .superconductor/config.json when present, root manifests, lockfiles, workspace configuration, and package scripts.
Choose commands only when repository evidence supports them:
- Derive setup from the committed package-manager lockfile, such as
pnpm install --frozen-lockfileornpm ci. - Prefer an existing
devorstartscript for the default run target. - Inspect the framework CLI help before adding host, port, or forwarding arguments. Do not guess flag syntax.
- For a browser-facing development server, configure an IPv4-reachable bind address when the framework supports it; verify the exact invocation. For example, Astro accepts
pnpm run dev --host 0.0.0.0, while other frameworks differ. - Preserve an existing valid setup or run command unless evidence shows it is wrong.
- Ask one focused question when the package manager, run script, or intended service is ambiguous.
Do not invent teardown, stop, or cleanup commands. Add them only when the repository already defines the required lifecycle behavior.
Step 3: Write Repo-Owned Configuration
Use repo scope so managed worktrees inherit the configuration. Use sc, not direct JSON replacement, because it preserves unknown fields and validates atomically.
sc project scripts set setup "<verified setup command>" --scope repo --json
sc project scripts set run devserver --default "<verified run command>" --scope repo --jsonIf no setup command is needed, do not add an empty one. If the repository already has multiple named run targets, preserve them and choose a descriptive name for the new target rather than replacing unrelated entries.
Step 4: Verify
- Validate the resulting config:
sc project scripts check --json- Run the setup command once when it is safe and non-destructive.
- Start the configured target through
sc project scripts run <name> --json. - Confirm
sc project scripts status --jsonreports it running. - When the service exposes HTTP and its port is known from evidence, request its IPv4 loopback URL and require a successful response.
- Stop the target after verification unless the user asked to leave it running, then verify it is idle.
- Inspect
.superconductor/config.json,git diff, andgit status --short.
A successful process launch without a reachable service is not sufficient verification for a browser-facing app.
Step 5: Report
Report:
Configured `.superconductor/config.json`:
- Setup: `<command>` or none
- Default run target: `<name>` — `<command>`
Verification:
- Config validation — pass/fail
- Setup — pass/skipped with reason
- Runtime state — running/failed
- Service check — URL and result, or not applicable
- Final runtime state — idle or intentionally left running
Files changed:
- `.superconductor/config.json`Leave the configuration uncommitted unless the user separately requests a commit.
In this repository14 other files it gives its agents.
- AGENTS.mdAGENTS.md
- commitSkill
- frontend-designSkill
- githubSkill
- pi-extension-developmentSkill
- planSkill
- reviewSkill
- scoutSkill
- session-readerSkill
- skill-creatorSkill
- superconductorSkill
- visualize-planSkill
- workerSkill
- write-todosSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
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.

