Akashic

Any repo becomes a Collection — record, sleeve, liner, and volume.

Type: Web portal & VS Code · Category: AI & Dev Tools

Managed SaaS for bringing existing code into AI collaboration — not a PM overlay. Connect a Claude Fable, ChatGPT or Gemini spike, Cursor project, or legacy monolith as a Collection; group repos into Libraries; span entities in a Workspace. The Ladder graph compiler builds a 7-layer product graph from connected repos, detects structural gaps (missing outcomes, orphaned nodes, cross-repo duplicates, redundant paths, stack layer holes, under-utilized personas, gate coverage holes, missing human interactions), evaluates Product Board health, and generates ordered tickets in contract-first dependency order — Terraform for products. Canonical Prefect mastering on ObVect runs full analysis with tickets terminal after understanding and restructuring. Multi-surface tickets become reasoning graphs — ordered DAGs of agent personas (Shiva clears structure, then Nyame, Sunan Ambu, Waaq, Amaterasu, Heimdall, Maat, Poseidon, Zeus, and more). Choose engagement goals and Mandate Mode to steer analysis and CI recommendations. Opt in from Settings when you want your Collection to contribute to the pooled corpus. Thirty-one Auth0 identity-provisioned personas — structured as agent personas across roles, Anu at the MCP boundary, Ishtar for portal analytics — plus ie the platform helper work from the web portal or VS Code on ObVect.

Highlights

Collection → Library → Workspace — One repo is a Collection (record · sleeve · liner · volume). Related repos share a Library. Cross-entity collaboration lives in a Workspace.
Ladder graph compiler — Builds a 7-layer product graph from connected repos, detects eight classes of structural gap (missing outcomes, orphaned nodes, cross-repo duplicates, redundant paths, stack layer holes, under-utilized personas, gate coverage holes, missing human interactions), evaluates Product Board health with cross-repo duplication and consistency analysis, and generates ordered tickets in contract-first dependency order — Terraform for products.
Reasoning graphs — Ticket content maps to an ordered DAG of agent personas — e.g. Nyame → Ukko → Amaterasu → Heimdall for schema + API + UI + tests. Mastering stages still land on Record / Liner / Volume.
One shared reasoning graph — Every agent runs in a multi-turn loop (ObVect wraps any model that is not already loop-capable) and all of them read and write one shared reasoning graph tied to the ticket + reconciliation system — so the whole team of agents reasons against a single source of truth. HITL approval gates are built into the graph and bound to RBAC: the right role signs off on a node before its agent action commits.
Confidence-gated HITL — Every agent decision carries a confidence score. High-confidence steps auto-approve; low-confidence steps pause for human review. Agents ask for help exactly when they should.
Global kill switch — One button stops every agent across the platform — every loop, dispatcher, sandbox, and assistant. Enforced at both the execution layer and the Idio PDP.
Unified explain endpoint — Trace any agent interaction back to its reasoning, audit trail, provenance, and cost in one API call.
Real-time agent control — Pause, resume, or cancel any running agent or loop from the portal dashboard.
Agent personas — Thirty-one Auth0 M2M identities (Olorun, Shiva, Tengri, Janus, Maat, Nyame, Mercury, Anu, Ishtar, Poseidon, Zeus, Amaterasu, Ukko, …) — scoped OAuth, MCP tools, Signed Agents audit trails. Amaterasu + Sunan Ambu own the tools; Waaq owns packages and images; Zeus owns Terraform / Terraspace / multi-cloud deploy (Poseidon hands off CI); Anu owns the MCP boundary; Ishtar owns portal analytics and Change History rollbacks; ie guides onboarding and Mandate Mode (not an Auth0 spawn pool).
Mandate Mode & engagement goals — Prototype → build → cleanup → scale → maintain, plus brownfield / greenfield / scratch personas and preset goals — steer phase weights, tickets, and CI packages.
VCS corpus discovery — GitHub and Bitbucket scans (corpus_search → quality_score → classify → exemplar_record) build lane-organized exemplars and package proposals the analysis engine reuses.
Opt-in corpus sharing — Collections stay private unless you opt in from Settings — then agent traces can join the pooled corpus that improves shared platform models (e.g. NUF-1).
Prompts, not voice-first — The liner captures IDE, agent, and chat prompts that shaped the repo — plus emit_skillfiles conventions agents inherit. Voice capture is optional.
Reconciliation gate — Per-file tickets with LLM acceptance criteria; semantic dedup against live PM boards across the Library before anything syncs.
Powered by ObVect — Every agent call, retrieval, and reasoning-graph execution flows through ObVect — the Virtuo SDK, OmniRoute model gateway, and Neo4j + Milvus retrieval that power the entire Colossal Capital ecosystem.

Features

Web portal — Mastering passes, reasoning graph composer, Prefect run history, engagement goal picker, RAG chat with citations, ticket matrix, PM connections, agent approvals, deployments, agent control dashboard, training cost summary.
VS Code extension — Same Workspace as the portal — connect Collections, run analysis, semantic search from the editor.
Knowledge graph & codemap — Neo4j graph explorer plus architecture fingerprint and ripple-risk codemap on every Collection.
Idio-native PM OAuth — Connect Linear, Jira, Asana, Monday, ClickUp, and GitHub Issues at the account level — bind boards to repos after connect.
Managed CI/CD path — Tekton pipelines for builds, scans, signing, and SBOMs — portal deploy flow targeting your cloud.
Change History — Versioned snapshots and point-in-time rollback for tenant state, reasoning maps, and analysis runs. Audit the past without rebuilding it.

Highlights​

Features​

Tags​

Highlights

Features

Tags