Signal Loom
github.com
What it is
Signal Loom is an open-source local skill package for turning research materials into evidence-traceable infographics. It accepts reports, notes, datasets, interviews, or existing infographics together with an intended audience and purpose. The default result is a semantic, responsive web infographic; an approved story can be reconstructed for carousels or other requested formats. Each project creates a Loomfile folder containing evidence, claims, story, visual plan, infographic, review records, and checkpoints. Its workflow runs through ORIENT, SHAPE, PLAN, BUILD, FINISH, DISTRIBUTE, and VERIFY. The repository documents Codex and Claude Code as host environments, with installation by cloning the repository into the host’s personal skills directory. Included Python tools use the standard library, make no network calls, and contain no telemetry. Signal Loom does not independently verify truth, refresh facts without authorization, approve its own work, or publish artifacts. Static checks cannot establish factual accuracy, accessibility, security, or professional fitness.
Spec sheet
- Purpose
- Signal Loom turns research, reports, notes, and data into evidence-traceable infographics.Read 10 Oct 2026 · github.com
- Maker
- The GitHub maker profile identifies Sam “stunspot” Walker as a prompt engineer and AI systems designer based in Illinois, USA.Read 10 Oct 2026 · github.com
- AI host privacy
- The configured AI host may transmit prompts and files according to its own configuration and terms.Read 10 Oct 2026 · github.com
- Audience and inputs
- It is designed for work based on reports, research notes, datasets, interviews, or existing infographics, with an intended audience and purpose supplied by the user.Read 10 Oct 2026 · github.com
- Evidence limits
- Signal Loom does not invent missing evidence or independently establish whether a claim is true.Read 10 Oct 2026 · github.com
- Host integrations
- Signal Loom supports Codex and Claude Code as host environments.Read 4 Oct 2026 · github.com
- Installation
- The project instructs users to clone the complete repository into the host's personal skills directory.Read 4 Oct 2026 · github.com
- Local operation
- The included Python tools use the standard library, make no network calls, and include no telemetry; Loomfiles stay where users create them.Read 10 Oct 2026 · github.com
- No autonomous publication
- Signal Loom does not independently establish whether claims are true, refresh current facts without authorization, approve its own work, or publish artifacts on its own.Read 4 Oct 2026 · github.com
- Output
- Its default output is a semantic, responsive web infographic, with other formats available when requested.Read 10 Oct 2026 · github.com
- Project format
- It produces a Loomfile project folder containing evidence, claims, story, visual plan, infographic, review records, and checkpoints.Read 4 Oct 2026 · github.com
- Project records
- It creates a resumable Loomfile containing source evidence, claims, story, visual plan, infographic, review records, and checkpoints.Read 10 Oct 2026 · github.com
- Publication boundary
- It does not publish anything on its own, and built, reviewed, approved for export, and published are distinct states.Read 10 Oct 2026 · github.com
- Security limits
- The HTML inspector is not a sanitizer, and the packager does not scan file contents for secrets.Read 10 Oct 2026 · github.com
- Security scope
- Signal Loom is a local skill package, not a hosted service, and it does not collect telemetry or publish artifacts.Read 4 Oct 2026 · github.com
- Supported hosts
- The repository documents installation for Codex and Claude Code.Read 10 Oct 2026 · github.com
- Validation limits
- Static validation can catch defined structural defects but cannot prove factual accuracy, accessible real-world use, security, or professional fitness.Read 4 Oct 2026 · github.com
- Vulnerability reporting
- The security policy says no response-time or embargo guarantee is promised for vulnerability reports.Read 10 Oct 2026 · github.com
- Workflow
- Its documented workflow is ORIENT, SHAPE, PLAN, BUILD, FINISH, DISTRIBUTE, and VERIFY.Read 10 Oct 2026 · github.com
Plans and prices
Strong points and limits
Strong points
- Open-source MIT-licensed package
- Evidence and claims stored in Loomfiles
- Semantic responsive web infographic output
- Works with Codex and Claude Code
- Python tools make no network calls
Limits
- Does not independently verify claims
- Does not publish artifacts on its own
- Static checks are not security assessments
- AI host may transmit files under its terms
Who it suits
It suits people who need a local, resumable process for building evidence-linked infographics from supplied material. Users of Codex or Claude Code are the documented host audience.
Verdict
Signal Loom structures infographic work around traceable evidence, staged production, and review records. It remains a local skill package: users provide the material and approvals, while factual, security, accessibility, and publication decisions remain outside its automated validation.
Where Specifiction places it
- Infographic Makers#24 of 267.5