Tasks - omni

ready blocked needs help all epics kb

t-807Refactor shared systemd helpers across agentd and deployer

t-795Fix typecheck.sh jq parsing for LocalRun rundeps

t-791sd: investigate low acceptance rate + verify idle A100s shut down

t-790Intent: Impl-vs-spec verification

t-785Persistent agent env drift can set AGENTD_AGENT_COMMAND=agent and break services

t-782Fix ambiguous Data.ByteString.Base64 import in Omni/Health/Intervals

t-780Design and implement mkSkill Nix helper + skill CLI

t-779Align Anthropic OAuth flow with pi-mono

t-771Notify owner on LLM provider failover

t-765workflows: add system_prompt to reviewer + integrator frontmatter; drop obsolete coder/overseer

t-762Pipeline scheduler: verify commits reach live before marking status=done (fixes orphan-commit bug)

t-761Port pi's system prompt to Omni/Agent coding agents

t-760Design compaction strategy for Omni/Agent coding agent

t-759Epic: Unify agent runner — stdin-driven persistent mode in Omni/Agent

t-758Fix Omni/Ipt.py build failure

t-755Research: unify trace compatibility for persistent sessions

t-752Skill drift: ui-preview example references removed agentd-studio path

t-749Bug: task CLI fails to read mixed-case enum values in tasks.db

t-748Bug: agentd service restart can hang in stop-sigterm for ~2m

t-738Intent Phase 4 follow-up: fix exec_cmd code type (Nat→Int)

t-737Intent Phase 4 follow-up: align :foreign ordering semantics (parser/docs/tests)

t-736Intent Phase 4 follow-up: include declForeign in hash/store persistence

t-735Intent Phase 4 follow-up: make :foreign python executable in codegen

t-734Intent: include regression cases in default test workflow

t-733Intent: resolve remaining LANG.md contract contradictions

t-732Intent: make query-holds KB scoping strict by default

t-731Phase 4: Health metrics at /ava/health

t-730Phase 3: Engagement audit at /ava/engagement

t-729Phase 3: Context inspector

t-728Phase 2: Memory browser at /ava/memories

t-727Phase 2: Message detail view at /ava/chats/:chatId/messages/:msgId

t-724Ava Audit Dashboard

t-721Calendar web viewer

t-717Quant Phase 2: Alpha + Portfolio + Web

t-708Quant Phase 1: Data + Signals + Eval

t-702Secrets not provisioned after reboot

t-700Web architecture improvements (epic)

t-699Set up SMTP secrets for fund-digest service

t-696Migrate fund queries from hledger CLI to hledger-lib Haskell API

t-695Agentd daemon pi.log writing is broken - handle closed immediately

t-694.7Send to Review: integrate agent work into live branch

t-694.6Worktree diff view and file browser

t-690Bug: ship/push may report manifest update without persisting

t-685Add torch, mamba-ssm, causal-conv1d to omnirepo Python nix deps

t-684Guard against web deploy rollback from divergent worktrees

t-683MVP 3: Diverse model review (devil's advocate)

t-682MVP 2: Single-model auto-spec refinement loop

t-681MVP 1: Executor GO/NO-GO gate

t-680MVP 0: Baseline metrics for task pipeline

t-679Epic: Automated Task Up-Speccing Pipeline

t-678Speculative context pre-fetching (JIT optimization)

t-677JIT context rehydration via request_context tool

t-675.5Multi-model LLM routing infrastructure

t-675.4MVP 3: Diverse multi-model review

t-675.3MVP 2: Single-model auto-spec loop

t-675.2MVP 1: Executor GO/NO-GO gate

t-675.1MVP 0: Baseline metrics instrumentation

t-675Multi-LLM Task Specification Pipeline

t-667General OAuth 2.0 integration for Telegram bot

t-666Agent auth path drift: pipeline/agentd uses ~/.local/share/agent but fresh creds are in ~/.pi/agent

t-665Agentd: avoid zombie 'running' status when container exits before terminal event

t-664Pipeline: don't move failed runs to Review

t-658Investigate mbsync duplicate UID errors in Maildir

t-655Task system: validate namespace exists in repo when setting it

t-653Broadcom (AVGO) 10-K deep dive

t-652Research: SRAM companies 10-K deep dives — post-Groq/NVDA inference thesis

t-651Research: Data center cooling companies 10-K deep dives

t-650Research: Arista Networks (ANET) 10-K deep dive

t-642Task web UI improvements

t-631Support ~/work as a second Pipeline domain

t-630Fix agentd buffet provider support (claude-code, codex)

t-622fund command - domain budget management

t-620DevBrowser fails without local Playwright browser binaries

t-602Benchmark memory system against LoCoMo eval

t-549.3Interactive step-through debugger

t-549Agent debugger for Haskell agent

t-538.3Socket message protocol and handler

t-538.2Ava dual-channel listener (Telegram + socket)

t-538Ava-Pi notification system

t-535Add user_profiles table for per-user preferences

t-533Pre-compaction memory flush

t-529Graph: Terminal UI for navigation and composition

t-522Graph: TUI event handling

t-521Graph: TUI rendering

t-520Graph: TUI state and navigation

t-513Agent Graph (epic)

t-488LLM Newsreader

t-486Enforce PromptIR token budgets in Sequential

t-485Wire AdaptiveContextConfig thresholds

t-481Review memory improvements and prompt IR

t-455Research: auto-detection for best-of-N parallel execution

t-401Prototype: LLM as universal semantic interpreter (temperature=0)

t-400Design: leverage pure cognitive compute in Bayesian prompt framework

t-399Research: rate-distortion theory for prompt compression

t-397Research: static prompt analysis using open model internals

t-362Add capability-based permissions to agentd workflows

t-279Automated process review and improvement

t-265.7Drip scheduler for follow-up campaigns