Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add Fredasterehub/kiln/plugin install kilnWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/fredasterehub/kiln/doctor)<a href="https://agentmods.dev/commands/fredasterehub/kiln/doctor"><img src="https://agentmods.dev/badge/commands/fredasterehub/kiln/doctor/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/fredasterehub/kiln/doctor"><img src="https://agentmods.dev/badge/commands/fredasterehub/kiln/doctor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00024 | $0.01312 |
| Opus 5 | $0.00012 | $0.00656 |
| Sonnet 5 | $0.00005 | $0.00262 |
| Haiku 4.5 | $0.00002 | $0.00131 |
Grade A, and why
doctor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Take Kiln's pulse and report it. Two parts, nothing else in this turn: run the block below once, then render the report from what it printed.
This check writes nothing, anywhere — no .kiln/ tree, no state, no ledger, no temp file, no edit to any file. It never calls a model, never opens an interactive login, and never starts a build: codex login status reads credential state from disk and exits, and detect.js and the wake's --pulse view only read. The preflight that records evidence is a different run, later, inside the build session. Nothing observed here is described as recorded.
V="$(claude --version 2>/dev/null)"; E=$?; V="$(printf '%s\n' "$V" | head -1)"
if [ "$E" -eq 0 ] && [ -n "$V" ]; then echo "CLAUDE=$V"
elif command -v claude >/dev/null 2>&1; then echo "CLAUDE=mute"
else echo "CLAUDE=absent"; fi
V="$(node --version 2>/dev/null)"; E=$?
if [ "$E" -eq 0 ] && [ -n "$V" ]; then echo "NODE=$V"; else echo "NODE=absent"; fi
if command -v codex >/dev/null 2>&1; then
if codex login status </dev/null >/dev/null 2>&1; then echo "CODEX=signed-in"; else echo "CODEX=signed-out"; fi
else echo "CODEX=absent"; fi
if git rev-parse --is-inside-work-tree >/dev/null 2>&1; then echo "GIT=repo"; else echo "GIT=no-repo"; fi
node "${CLAUDE_PLUGIN_ROOT}/machine/src/detect.js" detect "${CLAUDE_PROJECT_DIR}" || echo "DETECT=did-not-run"
if [ -d "${CLAUDE_PROJECT_DIR}/.kiln" ]; then
node "${CLAUDE_PLUGIN_ROOT}/machine/src/wake.js" view "${CLAUDE_PROJECT_DIR}" --pulse || echo "PULSE=did-not-run"
fi
Render the report. Read ${CLAUDE_PLUGIN_ROOT}/product/voice/kiln.md first and speak every line under it — first person, the fact before the flourish, no emoji, no preamble, no closing offer. One line per item, in this order:
- The
claudeCLI — the version it answered with, or that it did not answer on my path. My floor is v2.1.198 — the oldest build that carries everything I use; if the version it gave is below that floor, say so and name the floor. - Node — the version, or absent. Absent is a fault: my scripts have no hands without it.
- Codex — installed and signed in, installed but signed out, or not installed at all. Never a fault: without it I build single-family, Claude minds checking each other, with the cost named in the same breath — two families are wrong in different ways, one family is not.
- The
Workflowtool — read from this session's own resolved tool-schema block, never from a shell probe. Present; absent, in which case phase and chunk sequencing runs through plain Node scripts instead; or unknown, when this host exposes no such block to the session at all. Unknown is reported as unknown, never rounded down to absent. - Agent teams — its own separate read of that same block, never inferred from what item 4 found. If they are off, name the switch:
CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1, set in the environment or in Claude Code settings, and the session restarted before it takes. - Git — a repository here, or none yet. Advisory: the anvil commits chunk by chunk and will want one eventually, never at this door.
- Where this project stands — the
modedetect.jsprinted, in plain words, adding nothing its output does not say. If it did not run, say plainly what stopped it and that this project went unexamined. - The pulse — only when a
.kiln/tree stands here and the--pulseview printed: its lines rendered one item each, in its own order — the phase and where the build stands in it; the last stop, any waiting operator question repeated verbatim, word for word; how many entries the build's memory holds. Adding nothing the pulse did not print. If the tree exists and the pulse did not run, say so and say the project's standing past its classification went unexamined. A project without a.kiln/tree ends the report at item 7 — its classification already says there is no pulse to take.
Mark a check that holds ✓ and a check that fails ✗. A fact that is neither — Codex absent, a capability switched off, no repository yet — carries no glyph and gets a plain sentence naming what changes. The pulse's lines are facts, never checks: no glyph on any of them, nothing observed there was recorded, and this door never adopts the build's persona and never builds — it looks, and says what it saw. Emoji never.
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.
- 9d ago First seen · 47 lines · 24 tokens per session scan A 14dc6b16be83
doctor is a command published in the GitHub repository Fredasterehub/kiln (222 stars, last pushed 1mo ago), licensed MIT. It adds 24 tokens to every session and 1,312 once invoked, about $0.0001 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 2026-08-30.
Other commands, from other repositories
cleanup
Systematically clean up code, remove dead code, and optimize project structure.
toh-fix
Evidence-first debugging — find the root cause with proof before touching code.
bug
Report a bug from a plain description — triggers investigation, structured documentation and fix coordination.
deep-audit.skeleton
Aciklama: Bu bolum Bootstrap tarafindan manifest verileriyle doldurulur. Gerekli manifest alanlari: project.description, stack.primary, project.structure, project.subprojects, stack.orm, stack.authmethod Ornek cikti.
fix-issues
Pick up one approved task, implement the fix, and close it.
security-audit
Full security review — GitHub alerts, static code analysis, live app check — then a task per finding.