Template for AI-assisted research writing where every claim traces to a checkable source and nothing is called done until it clears a readiness gate.
-
Updated
Aug 24, 2026 - Python
Template for AI-assisted research writing where every claim traces to a checkable source and nothing is called done until it clears a readiness gate.
Palimnote: upload a difficult scholarly text and get a traceable AI-assisted critical edition with resolved citations, reading routes and a knowledge graph.
GitHub Actions CI for pnpm monorepos covering three gotchas: pnpm/action-setup ordering, a pgvector Postgres service, and a CI-safe vs manual E2E split.
Fill-in handoff prompt for moving work between AI agent sessions: provenance header, verbatim sourced values, reverted-attempt disclosure, integration gate.
Pre-launch review gate: claim accuracy, accessibility, privacy, and synthetic demo-data disclosure
Render Markdown to PDF with pandoc and xelatex, then verify the output (page count, text overlap, hbox warnings) instead of trusting a zero exit code.
Dependency-free Node script that turns a phases-and-status JSON file into a static HTML status page, with a --check mode that fails CI when it goes stale.
Audit an outdated plan against current evidence with a REUSE, REVISE, REJECT, DEFER, VERIFY matrix
A gate lifecycle for multi-agent work: structured verdicts, failure routing, escalation, no self-approval
Catalog of small developer tools: pandoc table filter, .env secret prompt, migration preflight gate, pnpm CI template, status pages and a PDF verify kit.
An adversarial reviewer role: severity taxonomy, evidence citations, disconfirmation duty
Pandoc Lua filter that sizes Markdown pipe-table columns by content, so prose columns stop overflowing and no column shrinks below its longest word.
Catalog of process templates: an incident runbook and a launch-campaign scaffold with default-deny claims and rights ledgers.
Mine candidate references and idea leads from source text into deduplicated follow-up tables
Append-only conventions for decision logs, task ledgers, and open-question files in long-running projects
Paragraph-boundary-aware text chunking for fitting long documents into LLM context windows
A fixed briefing schema for specialist agents: scope, inputs, tools, output contract, acceptance tests
Verify quoted excerpts against source texts, tolerant of PDF line-wraps, Unicode punctuation, and elision
A structured cost model for AI-assisted projects: category taxonomy across low, base, and high scenarios
Detect AI-sounding patterns in prose and rewrite in an authentic human voice, with channel-specific rules
To associate your repository with the palimnote topic, visit your repo's landing page and select "manage topics."