Stunspot · Oct 27, 2026 · Oct 23, 2026
diff --git a/.gitattributes b/.gitattributes
@@ -20,3 +20,6 @@
 *.icns binary
 verification/whole-v0.1.1/** -text
 verification/hesperos-capabilities/** -text
+
+verification/purpose-rebuild/** -text
+verification/github-delivery/** -text
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,8 +1,14 @@
-# 0.1.1
+# Changelog
 
-Distinct delivery identity for the prior same-version cue repair, complete current-source kit and recognizable native atlas entries. Original doctrine/case revision/graph semantics retained.
+## 0.2.0 — 2026-10-15
 
-# Changelog
+Adds a persistent question-first inquiry desk, ordinary source notes/documents, entity and typed relationship authoring, native-case import/export, and a direct projection into the existing full atlas. Retains original attachment bytes and unknown case fields.
+
+The October 23 GitHub delivery corrects current-version download links and kit paths, and clarifies the inquiry desk installation and saved-state workflow.
+
+## 0.1.1 — 2026-10-14
+
+Distinct delivery identity for the prior same-version cue repair, complete current-source kit and recognizable native atlas entries. Original doctrine/case revision/graph semantics retained.
 
 ## 0.1.0 — 2026-09-08
 

diff --git a/README.md b/README.md
@@ -1,5 +1,10 @@
 # RED THREADS — Conspiracy Unraveler
 
+## Start with your question
+
+Open the portable inquiry desk at `codex/red-threads/workspace/Open Red Threads.cmd` (Windows) or `Open Red Threads.command` (macOS) in the extracted release. Start an inquiry from a question, add a source note or ordinary document, name entities, and describe a typed relationship. The existing interactive atlas shows the result. You do not need to manufacture a case JSON first. Follow the [app workflow guide](src/red-threads/workspace/README.md) for saving, return, ordinary inputs and recovery.
+
+
 **Pull the thread. Find the mechanism.**
 
 ![RED THREADS investigative artwork: an investigator studies a network of people, institutions, and red connections](docs/assets/red-threads.png)
@@ -8,7 +13,7 @@ A grant. A board seat. A shared adviser. Three articles repeating one claim. Int
 
 RED THREADS gives your AI investigative craft for following money, resources, obligations, influence, and decision rights. It helps turn a hunch or a pile of records into a source-linked case, competing explanations, useful next leads, and an interactive atlas you can inspect offline.
 
-**[Explore the fictional demo](https://stunspot.github.io/red-threads/demo.html)** · **[Download v0.1.1](https://github.com/Stunspot/red-threads/releases/download/v0.1.1/red-threads-v0.1.1.zip)** · **[Install](documentation/installation.md)** · **[Run your first case](documentation/first-case.md)**
+**[Explore the fictional demo](https://stunspot.github.io/red-threads/demo.html)** · **[Download v0.2.0](https://github.com/Stunspot/red-threads/releases/download/v0.2.0/red-threads-v0.2.0.zip)** · **[Install](documentation/installation.md)** · **[Run your first case](documentation/first-case.md)**
 
 ## See what a connection actually carries
 

diff --git a/VERSION b/VERSION
@@ -1 +1 @@
-0.1.1
+0.2.0
diff --git a/distribution/RED THREADS Conspiracy Unraveler T-Free v0.2.0 Extra.md b/distribution/RED THREADS Conspiracy Unraveler T-Free v0.2.0 Extra.md
@@ -0,0 +1,41 @@
+# Description
+
+RED THREADS turns a suspicious connection into an investigation you can see, interrogate and pursue. Feed it a transcript, a document, an institutional hairball or a well-earned hunch. It follows money, personnel, obligations, access, infrastructure, narratives and decision rights, then helps you work out what those relationships actually explain. Who funds the activity? Who can approve the brief? Who can withhold the resource? What changes when the dates line up—or fail to?
+
+The interactive atlas makes the investigation a geometry. Explore the wider network, focus on one actor, inspect a relationship's evidence, follow a path, filter the time period, or temporarily remove a node to see which connections remain. Linked views expose the timeline, source ancestry, rival explanations, unanswered leads and actor strategy. The source behind an impressive claim stays within reach, including when twelve articles turn out to be one article wearing eleven hats.
+
+RED THREADS gives intuition room to work. A hunch becomes a hypothesis with consequences worth checking; strategy and game theory expose constraints, incentives, outside options and credible moves. Funding, agreement, coordination and command keep their distinct meanings. The next research move aims to reveal a mechanism or separate explanations. Record-finding playbooks turn a suspicion into an obtainable next document. Campaign checkpoints preserve coverage and unfinished questions; guarded case revisions and transitive dependency reports help carry corrections through hypotheses, games and leads while retaining independent support. The local tools produce an offline atlas, structural discovery prompts, ranked research leads and JSON/CSV exports without a hosted service. Your red-string board has acquired standards. It is still allowed to be interesting.
+
+# Open the inquiry desk
+
+Choose New inquiry, write your question, and add a source note or ordinary TXT/Markdown/DOCX/PDF/image. Read text and DOCX in the desk; retained PDF/image originals remain available for inspection. Name entities and describe a precise sourced relationship. The full existing atlas stays below the workbench, with a separate-tab option for exploration. Cases persist outside the installation and can be downloaded in the native format. The app never invents relationships from attachments.
+
+# Usage Notes
+
+Copy the .zip from "Additional Files" to your Codex, Claude Code, or other harness, or add the .zip to a project knowledge base in Chat. Attach or reference the .zip in chat and say, "Install this Augment." The prompt in this post is optional.
+
+This archive contains a complete Codex skill. In a chat without skill installation, use its material as reference. Other harnesses need their own compatible skill-loading route.
+
+- Begin with a question, transcript, URL, document, existing case or rough theory. State any research boundary or time budget that matters.
+- Ask for the atlas when a case becomes tangled, then inspect relationships, dates, sources, explanations and actor moves. The included Meridian case is entirely fictional and makes a useful first tour.
+- Python 3.10+ and a modern browser enable the generated offline atlas and case tools. Live evidence collection uses the browsing and file tools available in your host. No paid API or database is required.
+- Keep the current case JSON and evidence notes with your research files. Save/Load view preserves attention; resume restores the research frontier. Review source impact and use a new case revision when evidence changes. JSON/CSV export keeps records portable.
+- Treat path traces and node removal as ways to inspect connectivity. Ask for the missing mechanism or next discriminating record before turning a visual pattern into an assertion about control.
+- Draft articles and briefings when you want them. Publishing, outreach, paid data access and recurring monitoring remain actions you choose.
+
+The 0.2.0 ZIP contains its complete selected source kit, Codex runtime and nested Claude ZIP. The atlas has accepted-art tab/native icons and optional current-file desktop entries; follow the installation and native-entry guides. Mac execution remains unqualified.
+
+# Changelog
+
+v0.2.0 - Adds a persistent ordinary-input inquiry desk feeding the existing investigative atlas; retains native cases, source originals and unknown fields.
+
+
+v0.1.1 - Distinct delivery identity, source-complete rebuild kit and recognizable native atlas entries; investigative doctrine and case operations unchanged.
+
+v0.1.0 - Initial release: investigative mechanism tracing; record-finding and reconciliation playbooks; rival hypotheses and actor strategy; campaign coverage and resumption; complete declared dependency tracing; guarded case revisions; curated offline atlas with label-aware routing and Save/Load views; structural scans, lead frontiers, JSON/CSV export and fictional worked investigations.
+
+# Tags
+
+investigative research, influence networks, power mapping, open source intelligence, source analysis, knowledge graphs, visual analysis, hypothesis testing, game theory, institutional analysis, narrative analysis, Agent SKILL, Augment
+
+Support: [stunspot Prompting Discord](https://discord.gg/stunspot)
diff --git a/distribution/RED THREADS Conspiracy Unraveler T-Free v0.2.0.md b/distribution/RED THREADS Conspiracy Unraveler T-Free v0.2.0.md
@@ -0,0 +1,21 @@
+# RED THREADS Conspiracy Unraveler v0.2.0
+
+Attach the ZIP and say, "Install this Augment."
+
+Use this optional prompt with the attached archive:
+
+```text
+Install RED THREADS Conspiracy Unraveler from this ZIP using this host's supported skill installation route. The outer RED-THREADS-v0.2.0 source kit includes two host packages. For Codex, install the complete codex/red-threads folder. For Claude, import the inner claude/red-threads-v0.2.0.zip, which contains the complete red-threads skill. Each includes SKILL.md, references, knowledge, scripts, assets, examples and license notices. Keep the installed skill folder intact. Read SKILL.md and preserve the caller's identity; no additional persona or plugin server is required. Keep future investigative cases in my workspace, outside the installed skill. Verify the installed files and explain any host reload needed; distinguish file installation from successful skill discovery. If this chat cannot install skills, use the attached material as a reference and tell me that limitation plainly.
+
+Then help me begin with the lead or question I provide. Build a useful first map, follow the mechanisms, and keep my hunches available as working hypotheses. Use the bundled atlas when Python and a browser are available. Ask only for information that changes the investigation.
+Open the inquiry desk for direct work from my question and source material. Use its ordinary source intake and typed connections to make a useful first neighborhood visible, then inspect the full atlas. Preserve existing native case fields and source originals when returning to advanced research. Read workspace/README.md for save and recovery behavior.
+
+```
+
+After installation, start naturally: "Use RED THREADS to investigate this transcript and map the money, people, resources and decision rights behind it."
+
+For a quick tour, ask: "Render the included fictional Meridian case into my workspace and walk me through the source ancestry, dates and rival explanations."
+
+The generated atlas needs Python 3.10+ and a modern browser. Live investigation uses the host's available browsing and file tools; the atlas itself works offline. The method remains usable in conversation when interactive rendering is unavailable.
+
+If an output already exists, use a new explicit filename. A validation error identifies a case field or reference to correct before rendering. Keep the current case JSON, its evidence notes and the atlas together. Use Save view and Load view for presentation state. Ask RED THREADS to run resume when returning, and to inspect impact before applying a source correction to a new case revision. The case data and evidence let you revise the investigation later.
diff --git a/distribution/install-red-threads.md b/distribution/install-red-threads.md
@@ -1,13 +1,15 @@
-# RED THREADS Conspiracy Unraveler v0.1.1
+# RED THREADS Conspiracy Unraveler v0.2.0
 
 Attach the ZIP and say, "Install this Augment."
 
 Use this optional prompt with the attached archive:
 
 ```text
-Install RED THREADS Conspiracy Unraveler from this ZIP using this host's supported skill installation route. The outer RED-THREADS-v0.1.1 source kit includes two host packages. For Codex, install the complete codex/red-threads folder. For Claude, import the inner claude/red-threads-v0.1.1.zip, which contains the complete red-threads skill. Each includes SKILL.md, references, knowledge, scripts, assets, examples and license notices. Keep the installed skill folder intact. Read SKILL.md and preserve the caller's identity; no additional persona or plugin server is required. Keep future investigative cases in my workspace, outside the installed skill. Verify the installed files and explain any host reload needed; distinguish file installation from successful skill discovery. If this chat cannot install skills, use the attached material as a reference and tell me that limitation plainly.
+Install RED THREADS Conspiracy Unraveler from this ZIP using this host's supported skill installation route. The outer RED-THREADS-v0.2.0 source kit includes two host packages. For Codex, install the complete codex/red-threads folder. For Claude, import the inner claude/red-threads-v0.2.0.zip, which contains the complete red-threads skill. Each includes SKILL.md, references, knowledge, scripts, assets, examples and license notices. Keep the installed skill folder intact. Read SKILL.md and preserve the caller's identity; no additional persona or plugin server is required. Keep future investigative cases in my workspace, outside the installed skill. Verify the installed files and explain any host reload needed; distinguish file installation from successful skill discovery. If this chat cannot install skills, use the attached material as a reference and tell me that limitation plainly.
 
 Then help me begin with the lead or question I provide. Build a useful first map, follow the mechanisms, and keep my hunches available as working hypotheses. Use the bundled atlas when Python and a browser are available. Ask only for information that changes the investigation.
+Open the inquiry desk for direct work from my question and source material. Use its ordinary source intake and typed connections to make a useful first neighborhood visible, then inspect the full atlas. Preserve existing native case fields and source originals when returning to advanced research. Read workspace/README.md for save and recovery behavior.
+
 ```
 
 After installation, start naturally: "Use RED THREADS to investigate this transcript and map the money, people, resources and decision rights behind it."

diff --git a/distribution/red-threads-companion.md b/distribution/red-threads-companion.md
@@ -6,6 +6,10 @@ The interactive atlas makes the investigation a geometry. Explore the wider netw
 
 RED THREADS gives intuition room to work. A hunch becomes a hypothesis with consequences worth checking; strategy and game theory expose constraints, incentives, outside options and credible moves. Funding, agreement, coordination and command keep their distinct meanings. The next research move aims to reveal a mechanism or separate explanations. Record-finding playbooks turn a suspicion into an obtainable next document. Campaign checkpoints preserve coverage and unfinished questions; guarded case revisions and transitive dependency reports help carry corrections through hypotheses, games and leads while retaining independent support. The local tools produce an offline atlas, structural discovery prompts, ranked research leads and JSON/CSV exports without a hosted service. Your red-string board has acquired standards. It is still allowed to be interesting.
 
+# Open the inquiry desk
+
+Choose New inquiry, write your question, and add a source note or ordinary TXT/Markdown/DOCX/PDF/image. Read text and DOCX in the desk; retained PDF/image originals remain available for inspection. Name entities and describe a precise sourced relationship. The full existing atlas stays below the workbench, with a separate-tab option for exploration. Cases persist outside the installation and can be downloaded in the native format. The app never invents relationships from attachments.
+
 # Usage Notes
 
 Copy the .zip from "Additional Files" to your Codex, Claude Code, or other harness, or add the .zip to a project knowledge base in Chat. Attach or reference the .zip in chat and say, "Install this Augment." The prompt in this post is optional.
@@ -19,10 +23,13 @@ This archive contains a complete Codex skill. In a chat without skill installati
 - Treat path traces and node removal as ways to inspect connectivity. Ask for the missing mechanism or next discriminating record before turning a visual pattern into an assertion about control.
 - Draft articles and briefings when you want them. Publishing, outreach, paid data access and recurring monitoring remain actions you choose.
 
-The0.1.1 ZIP contains its complete selected source kit, Codex runtime and nested Claude ZIP. The atlas has accepted-art tab/native icons and optional current-file desktop entries; follow the installation and native-entry guides. Mac execution remains unqualified.
+The 0.2.0 ZIP contains its complete selected source kit, Codex runtime and nested Claude ZIP. The atlas has accepted-art tab/native icons and optional current-file desktop entries; follow the installation and native-entry guides. Mac execution remains unqualified.
 
 # Changelog
 
+v0.2.0 - Adds a persistent ordinary-input inquiry desk feeding the existing investigative atlas; retains native cases, source originals and unknown fields.
+
+
 v0.1.1 - Distinct delivery identity, source-complete rebuild kit and recognizable native atlas entries; investigative doctrine and case operations unchanged.
 
 v0.1.0 - Initial release: investigative mechanism tracing; record-finding and reconciliation playbooks; rival hypotheses and actor strategy; campaign coverage and resumption; complete declared dependency tracing; guarded case revisions; curated offline atlas with label-aware routing and Save/Load views; structural scans, lead frontiers, JSON/CSV export and fictional worked investigations.