Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/gaimsdevsoftware/glaze-coder/glazegit clone --depth 1 https://github.com/GaimsDevSoftware/glaze-coderWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00032 | $0.00391 |
| Opus 5 | $0.00016 | $0.00196 |
| Sonnet 5 | $0.00006 | $0.00078 |
| Haiku 4.5 | $0.00003 | $0.00039 |
Grade A, and why
glaze scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Use the glaze-app-dev skill to help the user develop a Glaze app for free (with
their own Claude subscription, not Glaze's paid agent).
Argument (optional): $ARGUMENTS = the app the user wants to work on, and/or the change
they want to make.
Steps:
- Detect the Glaze install and set the bundled Node on PATH (see the skill, step 0).
- If no app was named, run the equivalent of
ls "$GLAZE_BASE/apps"and read each app's.glaze-sources/package.jsonto list apps by product name, then ask which one. - Read the version-matched guidance for that app before coding: the app's
.glaze-sources/CLAUDE.md,.glaze_memory/PROJECT-CONTEXT.md, the relevant skill in"$GLAZE_BASE/agent-resources/current/.claude/skills/", andGLAZE-APP-GUIDE.mdas needed. - Make surgical edits only inside
.glaze-sources/, honoring every constraint in the skill (never touch.glaze/,@glaze/core, or the managed.npmrc; no CSS blur backgrounds; forbidden imports). - Validate with
npm run type-check, thennpm run build, thenopen ../*.app(ornpm run devfor live iteration), all local, all free. - Tell the user manual edits aren't captured in Glaze's version history; if they want a checkpoint, they can make one small prompt to Glaze's agent afterward.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 27 lines · 32 tokens per session scan A 94ebb17955af
glaze is a command published in the GitHub repository GaimsDevSoftware/glaze-coder (8 stars, last pushed 12d ago), licensed MIT. It adds 32 tokens to every session and 391 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
app
Control macOS apps via peekaboo app.
mcp
Run Peekaboo as an MCP server via peekaboo mcp.
move
Position the cursor via peekaboo move.
set-value
Set accessibility element values directly via peekaboo set-value.
spec-design
Create comprehensive technical design for a specification.
vibe-agents
Step 4 of the vibe-coding workflow: generate AGENTS.md + tool configs so the AI builder stays on track.