squidsquad

AI Community v0.43.0

Skill notes

Roles

SquidSquad always has PM, QA, and DM agents. Dev agents are flexible — you define them at setup time. Agent · Owns · Loop [role] Lead (one per dev role) · Code for that role, bugs and features via GitHub Issues · Ralph Loop (fix bugs → implement features → test → push) QA · Test results, qa/qa-log.md, bug verification, feature testing · Ralph Loop (E2E tests → verify bugs → test features → health checks → push)

File Structure Generated

When you invoke SquidSquad, it creates the following inside your project root. One folder is generated per dev agent — the example below shows a be-only setup: Note: All agents share a single GitHub Issues tracker on the repo. Issues are labeled by role, type, and status. QA queries for Pending Test items, verifies them, and comments. DM queries for Pending Ship items and handles delivery. For fe, be the structure gains a fe/ folder alongside be/.

Tracker — GitHub Issues

SquidSquad uses GitHub Issues as its tracker. All bugs and features are GitHub Issues with structured labels. Agents use gh CLI to create, query, update, and comment.

Label Taxonomy

Category · Labels · Purpose Type · type:issue, type:task · What kind of item Priority · priority:high, priority:medium, priority:low · Triage ordering Status · status:open, status:pending, status:planning, status:planned, status:approved, status:in-progress, status:pending-test, status:pending-human-review, status:pending-human-setup, status:pending-ship, status:shipped · Workflow state

Bug Flow

Status labels: status:open → status:in-progress → status:pending-test → status:pending-ship → (Issue closed)

Fluxly profile Runtime and permissions are inferred by Fluxly static checks, not author declarations or runtime test results.
Domain
AI
Author-declared agents
No explicit declaration found; this is not inferred or tested compatibility
Static check
98 / 100 · heuristic scan, not runtime safety proof
Runtime evidence
Not run; no isolated runtime test evidence is currently recorded
Author / version / license
@WallyDoodlez · v0.43.0 · AGPL-3.0
Token usage
Heavy
Setup complexity
Guided setup
External API key
Not required
Operating systems
macOS · Linux · Windows
Runtime requirements
Python
Detected file/system behavior
  • Read-only
  • Write / modify
  • Shell exec
Network behavior
Local-only
Install commands
None (reference only)

Heads up: Tool permissions are not declared; actual access depends on the runtime.

01
Step 1 — Detect Version Gap

Read `.squidsquad/config.md` to get installed `SquidSquad Version`. Read `SKILL.md` frontmatter for current version. If both match: tell the user they're up to date and stop.

02
Step 2 — Read Install Spec (if present)

Read `.squidsquad/.install-spec.json` for the install-time configuration (agent list, preset, flags). If the file does not exist (pre-install-spec installations), derive the agent list from the `Dev Agents` field in `config.md` and proceed — the upgrade flow works without it.

03
Step 3 — Regenerate Agent Templates

Rebuild all agent CLAUDE.md files from the current sub-skill sources using the compose skill: Or directly via CLI: `python references/scripts/compose.py deploy-all` This regenerates `.squidsquad/[role]/CLAUDE.md` for every configured agent (dev agents from config + PM + QA + DM — all always present). Placeholder substitution (`[ROLE]`, `[INTERVAL]`, `[ROLE_TEST_CMD]`, etc.) is handled automatically by compose.py.

04
Step 4 — Patch Config Schema (v1 → v2)

If `Architecture Version` in config.md is `1` (or absent), add any missing v2 sections with defaults. Do not delete existing v1 sections — agents still read them via config.py. Check for and add these sections if missing (with defaults): `## Preset` — `Id: software-dev` `## Tools` — `(none)` `## Loop` — `Interval Minutes: [existing interval value]`, `Context Threshold: [existing threshold value]`

05
Step 5 — Sync Labels

This is idempotent — it creates any missing GitHub Issue labels and skips existing ones.

Discussion

Powered by GitHub Discussions. Sign in with GitHub to comment, react, or subscribe.