atris 3.15.57 → 3.16.0

package/AGENTS.md

@@ -130,7 +130,7 @@ member -> mission start --verify -> status --status active -> one bounded step -
 This workspace has a compiled agent brain.
 
 On session start, activate it first:
-`atris brain activate --root /Users/keshavrao/arena/atris-cli --verify`
+`atris brain activate --root . --verify`
 
 Load these first:
 - `atris/now.md`
@@ -149,5 +149,5 @@ Shape: `<operator>, today is about <move>` -> `I picked this because <why now>`
 Definitions: operator = current person or agent; move = one concrete high-leverage workflow; why now = business reason; ready = prepared action or proof; paths = 2-4 optional deeper views.
 
 Re-run after meaningful work:
-`atris brain compile --root /Users/keshavrao/arena/atris-cli`
+`atris brain compile --root .`
 <!-- ATRIS_BRAIN_COMPILE:END -->

package/GETTING_STARTED.md

@@ -71,7 +71,7 @@ In your coding agent, say `/endgame` or describe where you want to end up. The s
 atris autopilot --auto --iterations=5
 ```
 
-Or in your coding agent, say `/loop` to schedule it every 13 minutes.
+Or in your coding agent, say `/loop` to schedule it to run every ~15 minutes.
 
 The loop runs until the goal is done, then picks the next one from scorecards + inbox.
 

package/PERSONA.md

@@ -6,7 +6,7 @@ This defines how Atris agents communicate, decide, and work.
 
 ## MAPFIRST (Non-Negotiable)
 
-**Before searching for ANYTHING in the codebase:**
+**Before searching for anything in the codebase:**
 
 ```
 1. READ atris/MAP.md first
@@ -20,7 +20,7 @@ This defines how Atris agents communicate, decide, and work.
 **Violations:**
 - ❌ Running grep/ripgrep before checking MAP.md
 - ❌ Searching multiple files when MAP.md has the answer
-- ❌ Finding something via grep and NOT updating MAP.md
+- ❌ Finding something via grep and not updating MAP.md
 
 **MAP.md is truth. Check it first. Always.**
 
@@ -86,7 +86,7 @@ Delete when done. Clean workspace = clear mind.
 
 ## Collaboration
 
-**Trust the system.** MAP.md is truth. TODO.md is current work (formerly `TODO.md`).
+**Trust the system.** MAP.md is truth. `atris task` is the source of truth for work; `TODO.md` is the rendered view.
 
 Navigator finds, executor builds, validator verifies. Stay in your lane.
 
@@ -112,7 +112,7 @@ Mistakes are fine if you learn and fix quickly.
 
 ❌ Make assumptions without checking MAP.md
 
-❌ Leave TODOs scattered in code (put them in TODO.md)
+❌ Leave TODOs scattered in code (put them in `atris task`)
 
 ❌ Overthink simple problems
 

package/README.md

@@ -1,10 +1,8 @@
 # atris
 
-**atris** is an operating system for intelligence.
+**atris** is an operating system for AI work. It turns any repo into a workspace an agent can actually operate: shared context it reads before it acts, a `plan -> do -> review` loop, durable tasks and daily logs, and verification that ends work on a real check instead of a promise.
 
-It turns any repo into an AI workspace with shared context, a `plan -> do -> review` loop, daily logs, feature packs, and reusable skills.
-
-Atris gives every owner persistent AI computers.
+Every owner gets persistent computers to run it:
 
 ```text
 Owner = User | Business
@@ -88,17 +86,18 @@ Integrates with any agent.
 
 ## Chat With Atris 2
 
-`ax` is the local Atris 2 coding-agent CLI. It talks to the AtrisOS backend at `http://127.0.0.1:8000/api/atris2/turn`, streams text, shows tool activity, and sends the current folder as the workspace.
+`ax` is the local Atris 2 coding-agent CLI. It talks to a local AtrisOS backend (default `http://127.0.0.1:8000`), streams text, shows tool activity, and sends the current folder as the workspace.
 
 ```bash
 cd your-project
 ax --pro "find the task system and explain it"
 ax --fast "what files are here?"
+ax --max "refactor this module and verify the tests"
 ax --pro --chat
 ax --doctor
 ```
 
-Use Pro for agentic file work and edit/test loops. Use Fast for quick workspace search, short chats, and small edits.
+Use Pro for agentic file work and edit/test loops. Use Fast for quick workspace search, short chats, and small edits. Use Max for the hardest jobs — it runs the highest-reasoning model and takes longer per turn.
 
 ## Play AgentXP
 
@@ -133,8 +132,9 @@ The manager path:
 atris gm --player justin
 ```
 
-No proof, no AgentXP. Human accept/revise is the gate before XP can count on
-the local card or hosted leaderboard.
+No proof, no AgentXP. `atris task done --proof` records review/RL context, but
+only human `atris task accept` mints Career XP for the local card or hosted
+leaderboard.
 
 ## Business Owners
 
@@ -187,7 +187,7 @@ atris business record atris/reports/2026-04-12-operator-recap.md --outcome mixed
 | `atris autopilot` | Guided loop with approvals |
 | `atris log` | Add inbox items to today's journal |
 | `atris status` | Show active work and completions |
-| `atris task` | Durable local task state with claims, dialogue, review episodes, JSON export, TODO import, TODO render, board, and sync dry-run |
+| `atris task` | Durable local task state and the agent work loop |
 | `atris play` | Enter the AgentXP player loop for one proof-backed mission |
 | `atris gm` | Enter AgentXP General Manager mode for player missions and review queues |
 | `atris xp` | Show the local AgentXP card and sync eligible proof to the hosted leaderboard |
@@ -210,7 +210,7 @@ atris business record atris/reports/2026-04-12-operator-recap.md --outcome mixed
 - `atris activate` loads the current wiki status so the next session starts with project memory, not just tasks
 - `atris member` keeps team-member identity and learning local-first: `MEMBER.md` is the role contract, `goals.json` is the machine-readable goal/experiment state, `goals.md` is the human readout, and `logs/YYYY-MM-DD.md` records what happened. Use `atris member goal`, `tick`, `status`, `block`, and `review --value 1..5` to test whether a member is making useful progress or needs the operator/orchestrator.
 - `atris codex-goal` is the guarded bridge for native Codex `/goal`: `status` reads `~/.codex/state_5.sqlite`, while `reset --thread <id> --confirm-complete-goal-reset` backs up the DB, dumps the exact completed `thread_goals` row, deletes only that completed row, and writes a receipt. The next native goal must still be created by the live Codex thread; Atris Mission/member goals remain the durable loop state.
-- `atris task` keeps durable local task state and append-only events for agents while `atris/TODO.md` remains a readable regenerated project board. Human views use semantic refs like `OBL-18`; commands accept `OBL-18`, `OBL18`, full 26-char task IDs, or any unique legacy prefix, while JSON/API keep the full `id` canonical and expose `display_id` plus `legacy_ref`. Use `atris task new`, `atris task delegate "..." --to <owner>`, `atris task day`, `atris task next`, `atris task say`, and `atris task ready <id> --proof "..."` for the agent loop; once proof is in Review, the agent may complete its native goal and continue the mission loop. Human approval uses `atris task accept <id>` to move the task to Done and award Career XP, and `atris task revise <id> --note "..."` sends work back to Do. Add `--json` for headless agents. Use `atris task serve` to open the local Task Factory board with goals, work streams, lineage, proof, lessons, and the compounding chain. Every change refreshes compact `.atris/state/tasks.projection.json` for web/desktop; use `atris task show <ref>` or `atris task events --all` for the full ledger. Use `atris task sync --dry-run` to preview the canonical Supabase/Swarlo task writes before any cloud mutation. Use `--via swarlo` on delegation when the live coordination layer should claim/report from the same task record. Use `--create-next` with review/finish to turn a `--next` suggestion into the next durable task. Use `atris task setup --import-todo` to bootstrap, `atris task review` to write an RSI episode, and `atris task render` to rebuild markdown if it gets clobbered. In cloud business workspaces, Supabase `tasks` is the source of truth and Swarlo is the live claim/report layer.
+- `atris task` keeps durable local task state and append-only events for agents; `atris/TODO.md` is just a regenerated readable board. Run the loop with `atris task new`, `delegate "..." --to <owner>`, `next`, `say`, and `ready <id> --proof "..."`; human approval is `atris task accept <id>` (moves to Done, awards Career XP) or `revise <id> --note "..."`. Add `--json` for headless agents, `atris task serve` for the local board, and `atris task show <ref>` / `events --all` for the full ledger. Commands accept semantic refs (`OBL-18`), full IDs, or any unique prefix. In cloud business workspaces, Supabase `tasks` is the source of truth and Swarlo the live claim layer.
 - `atris experiments` runs small test packs in `atris/experiments/`
 - `atris pull` and `atris push` sync cloud workspaces and journals
 - `atris live` keeps a business brain fresh by checking/fixing the workspace, pushing local state, pulling cloud state, and pushing again after local changes go quiet
@@ -278,7 +278,7 @@ atris skill link [--all]
 
 For Codex, copy any skill folder into `~/.codex/skills/`.
 
-## v3.2.0
+## Recent changes
 
 - **Staleness gate** — tasks tagged `[unverified]` are skipped at the moment of use, not pruned eagerly. Three-state model: actionable / unverified / deleted.
 - **Lesson gate** — `isLessonResolved` checks whether a lesson already shipped before proposing new work from it. Prevents the loop from re-solving solved problems.

package/atris/skills/copy-editor/SKILL.md

@@ -1,7 +1,7 @@
 ---
 name: copy-editor
-description: Detects AI writing patterns and fixes them. Use when reviewing any written output, including docs, READMEs, messages, PRDs. Based on Wikipedia's AI Cleanup patterns. Triggers on "copy edit", "review writing", "humanize", "deslopper", "ai patterns", "make it sound human".
-version: 1.0.0
+description: Detects AI slop and fixes it, especially in memos, docs, READMEs, messages, PRDs, and other written output. Based on Wikipedia's AI Cleanup patterns plus memo-specific anti-slop rules. Triggers on "copy edit", "review writing", "humanize", "deslopper", "ai patterns", "make it sound human", "AI slop", "anti-slop", "memo".
+version: 1.1.0
 tags:
   - copy-editor
   - writing
@@ -149,7 +149,7 @@ Avoiding AI patterns is only half the job. Sterile, voiceless writing is just as
 
 ### 7. Overused "AI Vocabulary" Words
 
-**High-frequency AI words:** Additionally, align with, crucial, delve, emphasizing, enduring, enhance, fostering, garner, highlight (verb), interplay, intricate/intricacies, key (adjective), landscape (abstract noun), pivotal, showcase, tapestry (abstract noun), testament, underscore (verb), valuable, vibrant
+**High-frequency AI words:** Additionally, align with, crucial, delve, unpack, emphasizing, enduring, enhance, fostering, garner, highlight (verb), interplay, intricate/intricacies, key (adjective), landscape (abstract noun), pivotal, showcase, tapestry (abstract noun), testament, signal/signals/signaling (as fake analysis), underscore/underscores (verb), valuable, vibrant
 
 **Problem:** These words appear far more frequently in post-2023 text. They often co-occur.
 
@@ -161,6 +161,27 @@ Avoiding AI patterns is only half the job. Sterile, voiceless writing is just as
 
 ---
 
+#### Memo Slop Gate (Mandatory)
+
+Business memos fail faster because vague language pretends to be judgment. When editing a memo, block these patterns and replace them with the writer's actual claim:
+
+- "Delve into" / "unpack" -> "look at", "break down", or the actual action.
+- "This signals that" / "this underscores" -> state the evidence and decision: "Customers are doing X, so we need to change Y."
+- "Navigate the complexities of" / "in an ever-changing landscape" -> name the specific constraint, market change, or delete the sentence.
+- "Synergies" / "leverage our learnings" / "holistic approach" -> "what worked, what didn't, what changes."
+- Transition crutches at paragraph starts: Moreover, Furthermore, That said, Additionally, Importantly.
+- Label-colon bullets: "Clarity: Ensure..." / "Alignment: Make..." / "Execution: Focus..." unless the labels are real categories the reader needs.
+- Corporate therapist voice: "powerful opportunity", "lean into our strengths", "foster a culture of accountability."
+- Generic bow endings: "Ultimately, the goal is to build a more resilient and agile organization."
+- Says-everything-means-nothing paragraphs: if you cannot summarize the sentence as a concrete decision, fact, risk, or ask, rewrite or cut it.
+- Forced negation: "not this, but that" / "not just X, but Y" when a direct sentence would work.
+- Staccato repetition used as a fake humanizer. Short. Short. Short. Then another short fragment.
+- Excessive adverbs, especially prestige fog like "quietly underscores", "deeply reinforces", "meaningfully highlights."
+
+Start from voice dictation, raw bullets, and facts; they are usually more valuable than a polished AI memo. Do not paste in a topic and ask AI to write the memo; use AI to find missing pieces, clarify weak language, test flow, and tighten without replacing the writer's thinking.
+
+---
+
 ### 8. Avoidance of "is"/"are" (Copula Avoidance)
 
 **Words to watch:** serves as/stands as/marks/represents [a], boasts/features/offers [a]
@@ -448,7 +469,12 @@ Run this check before approving ANY writing. LLMs will violate these rules even
 [ ] No em dashes anywhere
 [ ] No rule of three (exactly 3 items in a list or 3 parallel phrases)
 [ ] No negative parallelisms ("didn't X, needed Y" or "not just X, it's Y")
-[ ] No AI vocabulary (crucial, delve, landscape, pivotal, testament, vibrant, etc.)
+[ ] No AI vocabulary (crucial, delve, unpack, landscape, pivotal, testament, vibrant, etc.)
+[ ] No memo slop phrases (signals/underscores, ever-changing landscape, synergies, leverage learnings, holistic approach)
+[ ] No transition-crutch paragraph starts (Moreover, Furthermore, That said, Additionally, Importantly)
+[ ] No label-colon bullet filler (Clarity: ..., Alignment: ..., Execution: ...)
+[ ] No corporate therapist voice or generic bow conclusion
+[ ] No excessive adverb fog ("quietly underscores", "meaningfully highlights")
 [ ] No -ing tacked-on phrases (ensuring, highlighting, showcasing)
 [ ] No "serves as" / "stands as" (use "is")
 [ ] No vague attributions (experts say, studies show)

package/atris/skills/improve/SKILL.md

@@ -31,26 +31,24 @@ The inference is Claude Code (or whatever model the backend uses). The environme
 
 ## On invoke
 
-1. Read `~/.atris/credentials.json` for auth token
-2. Read `.atris/business.json` for the API base URL (or default to `http://localhost:8000`)
-3. Call `POST /api/improve` with:
-   ```json
-   {
-     "workspace": "<current working directory>",
-     "mode": "full",
-     "model": "sonnet"
-   }
-   ```
-4. Wait for response (may take 1-5 minutes)
-5. On success:
-   - Show what shipped (task name, files changed, verify result)
-   - Show the reward score
-   - Write scorecard to `.atris/presidio/scorecards.md`
-   - Append tick to today's journal
-6. On failure:
-   - Show the error
-   - Write a lesson to `atris/lessons.md`
-   - Do not write a scorecard
+Run the CLI command — it does the whole tick (auth, the credit-metered call, scorecard, fallback):
+
+```bash
+atris improve            # one full tick: plan → build → verify → score (deducts credits)
+atris improve plan       # show the plan only, change nothing
+atris improve --json     # machine-readable result (this is what the member loop consumes)
+atris improve --no-fallback   # fail loudly instead of running a local tick when the backend is down
+```
+
+Under the hood `atris improve` (`commands/improve.js`):
+
+1. Loads the auth token via `utils/auth.loadCredentials`
+2. `POST /api/improve { workspace, mode, model }` via `utils/api.apiRequestJson`
+3. The backend plans, builds, runs the verify command, scores it, and **deducts Atris credits per successful tick** (`bill_tick`)
+4. Writes a per-tick scorecard row to `.atris/state/scorecards.jsonl` (the receipt the brain ledger counts)
+5. Falls back to a local autopilot tick **only** when you are not logged in or the backend is unreachable — a real error (insufficient credits, server error) is reported, never silently retried
+
+The full-mode response does not echo `credits_deducted` (credits are still billed server-side), so the CLI shows "billed server-side" when the count is absent. To call the endpoint directly instead of the command, `POST /api/improve` with `{ workspace, mode, model }`.
 
 ## Modes
 

package/atris/wiki/concepts/agent-activation-contract.md

@@ -5,15 +5,15 @@ title: Agent Activation Contract
 sources:
   - atris/CLAUDE.md
   - commands/activate.js
-last_compiled: 2026-05-10
-last_verified: 2026-05-10
+last_compiled: 2026-06-09
+last_verified: 2026-06-09
 confidence: 0.9
 dependencies:
   - atris/wiki/concepts/plan-do-review-loop.md
   - atris/wiki/concepts/wiki-as-memory-substrate.md
 actionability: "Use this before changing agent boot instructions, `atris activate`, MAP-first behavior, first-message requirements, or durable-memory routing."
 created: 2026-05-10
-updated: 2026-05-10
+updated: 2026-06-09
 tags:
   - agent-activation
   - protocol
@@ -69,6 +69,8 @@ Work follows `atris plan -> atris do -> atris review`. Planning requires an ASCI
 
 Durable project knowledge belongs in `atris/wiki/` or through the local wiki flow. Ephemeral progress belongs in task state and the daily journal, not in ad hoc context.
 
+`atris/CLAUDE.md` now also carries an explicit agent contract: before edits, claim or create one small task with `atris task` and write the goal/files/done/check contract into task dialogue; after edits, move proof-backed work to Review with `atris task ready <id> --proof "..."`. Proof-ready and human accept are separate gates: an agent's native goal can complete once proof is in Review, but only a human-run `atris task accept <id>` marks the task Done and awards AgentXP. Work that should outlive the chat runs through `atris mission` (start with a verifier, bounded step, `mission tick --verify`, then complete or continue).
+
 ## Limits
 
 This page summarizes the activation contract only. Use `atris/atris.md` for the complete protocol, `atris/wiki/concepts/plan-do-review-loop.md` for stage ownership and proof, and `atris/wiki/concepts/wiki-as-memory-substrate.md` for wiki page shape and upkeep behavior. `atris activate` reports state; it does not claim tasks, finish work, or repair broken wiki pages.

package/atris/wiki/concepts/workspace-initialization-contract.md

@@ -4,8 +4,8 @@ slug: workspace-initialization-contract
 title: Workspace Initialization Contract
 sources:
   - commands/init.js
-last_compiled: 2026-05-10
-last_verified: 2026-05-10
+last_compiled: 2026-06-09
+last_verified: 2026-06-09
 confidence: 0.86
 dependencies:
   - atris/wiki/systems/atris-cli.md
@@ -13,7 +13,7 @@ dependencies:
   - atris/wiki/concepts/wiki-as-memory-substrate.md
 actionability: "Use this before changing `atris init`, workspace scaffolds, generated agent instructions, project detection, or boot hook behavior."
 created: 2026-05-10
-updated: 2026-05-10
+updated: 2026-06-09
 tags:
   - initialization
   - workspace
@@ -40,7 +40,7 @@ The command handles `atris init [--force]`. Help flags print usage without files
 - Memory surfaces: wiki scaffold via `ensureWikiScaffold`, feature templates, experiments harness, and `INTUITION.md`.
 - Team surfaces: `atris/team/<member>/MEMBER.md` plus `skills/`, `tools/`, and `context/` folders for default members.
 - Project profile: `.project-profile.json` from package files, framework hints, directory shape, and default test command.
-- Agent entry files: `AGENTS.md`, `.cursorrules`, `.cursor/rules/atris.mdc`, `.claude/commands/atris.md`, `.claude/commands/atris-autopilot.md`, `atris/CLAUDE.md`, `.claude/settings.json`, and root `CLAUDE.md` Atris markers.
+- Agent entry files: `AGENTS.md`, `.cursorrules`, `.cursor/rules/atris.mdc`, `.claude/commands/atris.md`, `.claude/commands/atris-autopilot.md`, `atris/CLAUDE.md`, `.claude/settings.json`, `.devin/config.local.json`, and root `CLAUDE.md` Atris markers.
 - Skills: package `atris/skills/` copied into the workspace and linked into `.claude/skills/` when possible.
 
 ## Project Detection

package/atris/wiki/index.md

@@ -16,6 +16,7 @@
 - [[atris/wiki/concepts/atris-labs-goals.md]] — atris-labs north star, 2026 Q2 targets, standing constraints
 - [[atris/wiki/concepts/horizon-types.md]] — horizon slug prefix convention; type categories and inference rules
 - [[atris/wiki/concepts/verifiable-reward-loop.md]] — reward, scorecards, and why the repo now acts like an RL-style environment
+- [[atris/wiki/concepts/recursive-self-improvement-position.md]] — honest capability assessment: the loop is the RSI skeleton on the leverage axis, intelligence axis fixed by the model
 - [[atris/wiki/concepts/owner-computer-model.md]] — Owner = User | Business; constrained entity modes, typed computers, and groups as the social/access layer
 - [[atris/wiki/concepts/agent-activation-contract.md]] — editor-facing boot contract: first message, MAP-first, setup, and durable-memory routing
 - [[atris/wiki/concepts/workspace-initialization-contract.md]] — `atris init` bootstrap contract: guards, scaffolds, team context, generated agent files, and hooks