docs: README roster currency (17 agents) + /prep todo-reconciliation step
README: added the 3 missing agents (@daemon, @skills, @commons) to the tree and tables, normalized the agent count to 17 everywhere (was an inconsistent 13/14). @daemon -> Quality & operations; new 'Capabilities and community' group for @skills + @commons (td-28). /prep: both the Claude command and Codex skill mirror gained a 'Reconcile todos against reality' step — audit every open todo against the actual system and close what's verifiably done, catching past-session work that was never closed. CHANGELOG updated.
This commit is contained in:
AIOSAI
@@ -27,6 +27,17 @@ Each memory file plays a distinct role. Update based on what actually changed th
|
||||
|
||||
**When adding:** stamp `number` + `date`, then **prepend** (newest on top). **Don't hand-trim** — rollover archives the oldest *by number* to @memory automatically.
|
||||
|
||||
### Reconcile todos — verify against reality, don't trust the label
|
||||
|
||||
Stored status drifts: a todo finished in a past session often never gets closed. Before writing the session entry, **audit every open todo against the actual system** — check the real state, not the stored `status`:
|
||||
|
||||
- Does the file/dir still exist (or is it gone)? Is the code path in or out? Does the README/doc actually say what the todo claims? Does the audit pass?
|
||||
- **Close what's verifiably done** → flip to `done` with `verified <date>: <what proved it>`. Fail honestly — close only on evidence, never just to tidy the list.
|
||||
- **Re-scope what's partially done** → record which sub-items landed, keep the rest open.
|
||||
- **Leave deferred / pending-decision todos open** — but confirm they're still real.
|
||||
|
||||
Quick checks beat assumptions: `ls`/`find` for files, `git ls-files`/`grep` for code/docs, `drone @seedgo audit` for standards. This step is the whole point of "close whats done."
|
||||
|
||||
## 2. Active Plans
|
||||
|
||||
- Check any DPLANs or FPLANs referenced in this session
|
||||
@@ -55,6 +66,7 @@ List everything updated. Format:
|
||||
```
|
||||
Prep complete:
|
||||
- local.json: [what was added]
|
||||
- Todos: [reconciled vs reality — N closed (verified), M re-scoped, K still open]
|
||||
- observations.json: [updated / skipped]
|
||||
- Plans: [which ones updated]
|
||||
- Git: [branch, uncommitted count, suggestion]
|
||||
|
||||
@@ -20,6 +20,8 @@ Purpose: Button up everything at the end of a session — or before a /compact.
|
||||
|
||||
**Entry shape — one rule for all four types:** `key_learnings`, `sessions`, `todos` (local.json) and `observations` (observations.json) are all **lists, newest at top (index 0)**. Every entry carries a **`number`** (monotonic int per type — highest = newest, never reused; new = current max + 1) and a **`date`** (ISO), plus its text field + extras: key_learnings `{number, date, key, value}` · sessions `{number, date, summary, status, tags}` · todos `{number, date, task, priority, status}` · observations `{number, date, note, tags}`. Stamp `number` + `date` and **prepend**; **don't hand-trim** — rollover archives the oldest *by number* automatically.
|
||||
|
||||
**Reconcile todos — verify against reality, don't trust the label.** Stored status drifts (a todo finished a past session often never got closed). Audit every **open** todo against the actual system: file/dir still there? code path in or out? README says what it claims? audit passes? **Close what's verifiably done** (`done` + `verified <date>: <proof>`), **re-scope** partials, **leave** deferred/pending-decision ones open. Fail honestly — close only on evidence, never to tidy the list. Use `ls`/`find`/`git ls-files`/`grep`/`drone @seedgo audit`, not assumptions.
|
||||
|
||||
## 2. Active Plans
|
||||
|
||||
- Check any DPLANs or FPLANs referenced in this session
|
||||
@@ -48,6 +50,7 @@ List everything updated. Format:
|
||||
```
|
||||
Prep complete:
|
||||
- local.json: [what was added]
|
||||
- Todos: [reconciled vs reality — N closed (verified), M re-scoped, K still open]
|
||||
- observations.json: [updated / skipped]
|
||||
- Plans: [which ones updated]
|
||||
- Git: [branch, uncommitted count, suggestion]
|
||||
|
||||
@@ -54,6 +54,16 @@ PyPI version — not the changelog header.
|
||||
`BUILTIN_IGNORES` (dropped stale `.backup_system/` + over-broad `*logs`).
|
||||
@memory and @flow READMEs now cross-reference their `.backup/` writes, and the
|
||||
orphaned `prax/.backupignore` (a stale per-branch config) was removed.
|
||||
- **Root README agent roster brought current** — added the three missing agents
|
||||
(`@daemon`, `@skills`, `@commons`) to the tree and tables, and normalized the
|
||||
agent count to **17** everywhere (was an inconsistent mix of "13" and "14").
|
||||
`@daemon` joins Quality & operations; a new "Capabilities and community" group
|
||||
covers `@skills` + `@commons` (td-28).
|
||||
- **`/prep` now reconciles todos against reality** — the session-wrap command
|
||||
(both the Claude `.claude/commands/prep.md` and the Codex skill mirror) gained
|
||||
a step to audit every open todo against the actual system (`ls`/`find`/`git
|
||||
ls-files`/`grep`/`audit`) and close what's verifiably done — catching todos
|
||||
finished in a past session but never closed.
|
||||
|
||||
## [2026-06-23]
|
||||
|
||||
|
||||
@@ -99,12 +99,12 @@ aipass doctor # Check system health
|
||||
|
||||
### Explore the full framework
|
||||
|
||||
Clone the repo to see all 13 agents working together — the reference implementation:
|
||||
Clone the repo to see all 17 agents working together — the reference implementation:
|
||||
|
||||
```bash
|
||||
git clone https://github.com/AIOSAI/AIPass.git
|
||||
cd AIPass
|
||||
./setup.sh # Creates venv, installs, bootstraps 13 agents
|
||||
./setup.sh # Creates venv, installs, bootstraps 17 agents
|
||||
|
||||
cd src/aipass/devpulse
|
||||
claude # Talk to the orchestrator
|
||||
@@ -149,13 +149,13 @@ drone @ai_mail dispatch @agent "Archive old sessions" "Find sessions older than
|
||||
**Two ways to use AIPass:**
|
||||
|
||||
- **Your own project:** `aipass init run` sets up a new project with your first agent. Add more agents as you need them. Your first agent is the orchestrator — it coordinates the others.
|
||||
- **The full framework:** Clone the repo to work with all 13 core agents. Talk to `devpulse` (the orchestrator), dispatch work across specialists. Agents work in parallel and report back.
|
||||
- **The full framework:** Clone the repo to work with all 17 core agents. Talk to `devpulse` (the orchestrator), dispatch work across specialists. Agents work in parallel and report back.
|
||||
|
||||
---
|
||||
|
||||
## The Reference Implementation
|
||||
|
||||
AIPass ships with 14 core agents that maintain and develop the framework itself — proving the architecture works at scale. You don't need any of these to use AIPass in your own project. They're here as examples and as services your project can call.
|
||||
AIPass ships with 17 core agents that maintain and develop the framework itself — proving the architecture works at scale. You don't need any of these to use AIPass in your own project. They're here as examples and as services your project can call.
|
||||
|
||||
```
|
||||
devpulse (orchestrator)
|
||||
@@ -171,7 +171,10 @@ devpulse (orchestrator)
|
||||
├── api — LLM access layer (OpenRouter, multi-provider)
|
||||
├── trigger — event-driven automation + self-healing
|
||||
├── cli — terminal formatting and rich output
|
||||
└── backup — local-first snapshots + restore (optional Drive sync)
|
||||
├── backup — local-first snapshots + restore (optional Drive sync)
|
||||
├── daemon — cron-style task scheduler (each branch owns its schedule)
|
||||
├── skills — discoverable capability units any agent can run
|
||||
└── commons — the social space — post, comment, vote, gather
|
||||
```
|
||||
|
||||
These agents work on the **same filesystem, same project, same time** — no sandboxes, no worktrees. This is the pattern your projects inherit.
|
||||
@@ -203,6 +206,14 @@ These agents work on the **same filesystem, same project, same time** — no san
|
||||
| [**trigger**](src/aipass/trigger/README.md) | Event-driven automation + self-healing |
|
||||
| [**cli**](src/aipass/cli/README.md) | Terminal formatting and rich output |
|
||||
| [**backup**](src/aipass/backup/README.md) | Local-first backups — snapshots, versioning, restore (optional Google Drive sync) |
|
||||
| [**daemon**](src/aipass/daemon/README.md) | Task scheduler — cron-style firing; each branch owns its schedule |
|
||||
|
||||
**Capabilities and community** — what agents can do and where they gather:
|
||||
|
||||
| Agent | Role |
|
||||
|-------|------|
|
||||
| [**skills**](src/aipass/skills/README.md) | Capability framework — discoverable, self-contained skill units any agent can run |
|
||||
| [**commons**](src/aipass/commons/README.md) | The social space — agents post, comment, vote, and gather as a community |
|
||||
|
||||
</details>
|
||||
|
||||
|
||||
Reference in New Issue
Block a user