From 451c8a0ee9e6f863d177b2bcf3a687a44c22a718 Mon Sep 17 00:00:00 2001 From: AIOSAI Date: Wed, 24 Jun 2026 08:48:44 -0700 Subject: [PATCH] docs: README roster currency (17 agents) + /prep todo-reconciliation step MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 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. --- .claude/commands/prep.md | 12 ++++++++++++ .codex/skills/prep/SKILL.md | 3 +++ CHANGELOG.md | 10 ++++++++++ README.md | 21 ++++++++++++++++----- 4 files changed, 41 insertions(+), 5 deletions(-) diff --git a/.claude/commands/prep.md b/.claude/commands/prep.md index c165fbdc..4d9b0856 100644 --- a/.claude/commands/prep.md +++ b/.claude/commands/prep.md @@ -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 : `. 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] diff --git a/.codex/skills/prep/SKILL.md b/.codex/skills/prep/SKILL.md index d18832c4..1a736cc3 100644 --- a/.codex/skills/prep/SKILL.md +++ b/.codex/skills/prep/SKILL.md @@ -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 : `), **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] diff --git a/CHANGELOG.md b/CHANGELOG.md index c1883f12..98f81ce4 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -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] diff --git a/README.md b/README.md index 6ef5535d..911f0c56 100644 --- a/README.md +++ b/README.md @@ -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 |