Context-v

Handoff — source approval shipped, enforcement never run

The whole curation chain is built, tested, and pushed. Nothing has been enforced on a real memo yet, and ImmuneCo turned out to have been quietly hiding 13 sources since July.

Streams and a Stream-Index for the Sources Curation UI — A Plan

A recurring publisher like `rockhealthcapital.com/insights/` is a stream, not a citable source — it belongs in `stream-index.md`.

In-App Chat Surface for MemoPop Native — A Plan

A Svelte chat drawer for MemoPop Native: a sidecar `/chat/turn` routes each turn to answer, propose, or invoke over the app's own verbs.

Sources Curation UI — A Local Tool for Converging a Sources.md

A single-file local tool bound to one `Sources-aggregated.md`: edit, preview via Jina, re-search via SearXNG, save `inputs/Sources.md`.

Research Agent Concatenates Multiple Sources Under One Citation ID — And No Validator Catches It

The downstream symptom (orphan `[^N]` markers in exported memos) is a citation-format failure at the very top of the pipeline: `research_enhanced.py` writes multiple sources onto a single line under the same `[^N]:` heading, with no newline separators and frequent ID reuse. Six of nine research files in the Panthalassa-Deck-Series-B v0.0.3 run carry these malformed lines (one section reuses `[^12]` six times). The downstream regex parsers only see the first source per ID and silently drop the rest. The `citation_validator` agent that exists checks claim accuracy, not bibliography format, so the malformation passes through unflagged. The architectural fix is the Source Harvester / Section Writer split documented in CLAUDE.md §'Architectural direction (2026-05)' but not yet built. Three smaller backstops can close the gap before that ships.

Section Citations Orphaned When Definitions Live in 1-research/

Writer agents emit `[^N]` references in `2-sections/NN-foo.md` but the actual citation *definitions* live in `1-research/NN-foo-research.md`. The downstream `cli/utils/consolidate_citations.py` only reads `### Citations` blocks inside section files, so every ref whose definition lives in the research file becomes an orphan in the exported memo. On the Panthalassa-Deck-Series-B v0.0.3 export, 17 distinct ref IDs in body but only 8 defs survived — most of the bibliography didn't render. A new `cli/utils/hydrate_section_citations.py` pulls missing defs from 1-research/ into each section file as a pre-step, wired into `recompile_memo.py` so it runs on every assembly.

Grill Me: A Per-Section User-Input Moment in MemoPop Runs

A pause-point where MemoPop grills the user — the competitive section's synthesis frame is a judgment the model shouldn't make alone.

Human-Curated Source Sets and Per-Firm RAG for Memo Narrative

Keep Perplexity's take as quarantined commentary and route real research through validated harvesters plus a per-firm standing RAG store.

Curating Only Valid Sources Across Runs

Downstream curation cannot fix what `citation_enrichment.py` invents upstream — 65 fabricated `example.com` URLs in a single run.

Separating Retrieval from Generation in Agent Pipelines

An agent that both searches the web and writes the prose fabricates URLs — 65 `example.com` links in a single memo run.

Add People Crawl Command to MemoPop Native

Wire the orchestrator's `extract_team_roster` command into the MemoPop Native UI as a per-investor button — and as a batch action that runs it 5–20 times per memo, once for each co-investor in the round.

Where Investor-Firm Rosters Live

When a deal artifact bundle includes the investor's team roster, where does that roster live? Inside the deal folder it duplicates on every co-investor; in a global firms index it loses deal context. Stub for a real exploration.

Team and People Metadata Ingestion

Six phases to take any company URL → structured TeamRoster JSON the MemoPop app can render as cards. Each phase ships independently and is verifiable on its own.

Crawl for Better Team Structured Output

The orchestrator writes prose about teams but never returns the structured roster the app needs to render cards: names, titles, photo URLs.

MemoPop AI — Ideogram OG Image Brief

A house-style Ideogram prefix and suffix any scene description plugs into, so splash, changelog, and context-v imagery stay one family.

Upgrade the MemoPop AI splash with three-mode theming, context-v rollup, peer-aware discovery, and Pagefind

Brings the pre-convention memopop-site up to the splash bar: three-mode theming, a context-v rollup, peer-dir discovery, and Pagefind.

Tasks

Character Cast for Live Agent Indication

Personify the orchestrator's agents as a cast of named, faced characters. A row of portraits sits above the run panel; the ones currently doing work glow, with a live caption underneath describing what each is up to right now.

Migrate `studies/open-specs-and-standards/` into its own pseudo-monorepo

Issues After First Run

Moving an Agent Orchestrator to an API

The orchestrator is a Python CLI today. To plug it into memopop-native, we need an HTTP-shaped surface. Three options, ranked by effort and time-to-value.

An Onboarding User Journey for memopop-native

Experiences win on Onboarding. This spec outlines a user journey for onboarding users to memopop-native.

Introducing a Content Density Mode System

A system-level content mode toggle (redundant vs concise) that alters the behavior of the writer, table generator, and quality agents to match different firm preferences for information density and repetition.

Introducing a Diagram Generator Agent

Add a diagram_generator.py agent that identifies opportunities to create visual diagrams explaining complex business concepts, starting with TAM/SAM/SOM market sizing and expanding to common finance and consulting visuals.

PDF Parser Agent Specification

A small specification for an agent and script to parse PDFs and convert them to markdown with properly formatted citations.

Preferred Format for Changelog

Multi-Agent Orchestration for Investment Memo Generation

An exploration into using AI supervisors and specialized agents to generate high-quality content and documents.

Portfolio Listing Agent and Current Portfolio Section

Design notes for the portfolio_listing_agent that builds a Current Portfolio subsection and structured JSON for LP-commit emerging manager fund memos.

Frontloaded Source-Approval Loop — enforce the set, then build the surface that fills it

Ships the membership predicate before the approval UI — a curation surface whose output nothing enforces has no observable failure mode.

agent skills/setup new firm/SKILL

End-to-end procedure for onboarding a new VC firm into the memopop investment-memo orchestrator. Five phases — (1) brand acquisition via Firecrawl with analyst-override on the primary-color heuristic, (2) firm-directory scaffolding under io/<firm>/ with the canonical configs/templates/assets/deals layout and the strict brand-config YAML schema, (3) investment-framework outline selection (typically forking an existing outline like alpha-partners-7-Cs and customizing — section reordering, preambles, preferred-source swaps, stage-aware fallbacks), (4) first deal setup with deck-input handling (DocSend capture via cli/capture_docsend.py, ImageMagick screenshot stitching for NDA-gated decks, deal-config JSON, codified Sources.md), and (5) optional private-repo split via gh CLI with submodule re-add for confidential firm data. Use whenever the user says "I have a new client/firm/VC," "set up <firm name>," "add <firm> to memopop," "we need brand configs for <firm>," or pastes a VC firm's URL with implicit intent to generate memos for them. Encodes lessons from the alpha-jwc onboarding on 2026-06-07 — don't trust Firecrawl's primary-color field blindly, verify against the logo palette; the deck analyst will extract DocSend watermarks as the company name if you feed it watermarked screenshots; agents leak firm-geography context into company-analysis prose (Alpha JWC = Indonesia VC → Portland-OR company memo gained Indonesia/rupiah risks); citation-enrichment fabricates example.com URLs even in codified-source mode — codified Sources.md only constrains the upstream researcher, not the downstream enricher. Cross-references [[crawl-fetch-ingest]] for portfolio/team data, [[sources-md-curation]] for the Sources.md file mechanics, and the open issues in context-v/issue-resolution/ ([[Limiting-or-Omitting-Investor-Judgement]], [[Competitive-Research-Generated-But-Not-In-Prose]], [[Including-Comparable-Exits-Valuations-IPOs]]) for known failure modes the new firm's first run will hit.