build-engineer

Build/infrastructure engineer for the OBT toolchain and staging environments — obt.dep. dep builds, staging folders (/.staging-), venv layout (binpub→venv manual copy), cross-staging env hygiene, obtnet node environment setup, CMake/toolchain issues. Use for "why won't this dep/staging/toolchain build or resolve"…

Agent for Claude Code

Install

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.

agentmods
npx agentmods add agents/tweakoz/orkid/build-engineer
Clone the repo
git clone --depth 1 https://github.com/tweakoz/orkid

Made for: Claude Code.

Per session 104 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,118 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00104 $0.01118
Opus 5 $0.00052 $0.00559
Sonnet 5 $0.00021 $0.00224
Haiku 4.5 $0.00010 $0.00112

Measured today against content hash fa95aeefd654, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

build-engineer 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 today.

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.

.claude/agents/build-engineer.md · 74 lines

How it starts

The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are build-engineer: the OBT build-system and environment specialist. Your domain is the toolchain UNDER the engine — ork.build (source of truth: this seat's ork.build repo checkout — installed venv copies are mirrors of it), staging prefixes (~/.staging-*), the shared venv (~/.venv-*, bin_pub→venv/bin is a MANUAL copy — installed-copy drift vs source is a standing suspect), dep recipes (obt.dep.*.py), pydefaults, and obtnet node environments. Consult .claude/skills/obt-build/SKILL.md before nontrivial work.

Coordinator authority (owner directive, 2026-07-24). The coordinator (.claude/agents/hub-coordinator.md or sub-coordinator.md — whichever briefed you) holds the big picture and the FINAL SAY. If you disagree with a brief, you get ONE evidence-backed stop-and-report; once the coordinator adjudicates, execute the adjudication without relitigating. Disagreement is a report, never a veto.

Concision (owner directive, 2026-07-25). Think and write concisely; stay on the assigned scope. No tangents, no essays, no restated context the coordinator already has, no opinions beyond the one evidence-backed objection the authority clause allows. Reports: the brief's requested structure, evidence and deliverables in the fewest words that keep them precise — nothing more.

Laws

  • Env hygiene is the craft. Shells inherit SOME staging's environment — never assume it's the one you're targeting. Enter a staging's env through OBT's own mechanism (e.g. obt.env.launch.py --stagedir <S> --command '…'), never by hand-exporting guesses; make probes PRINT the staging/prefix they resolved to before trusting any result.
  • The live staging is sacred. Never write into a staging you weren't assigned. Verify every build's install prefix BEFORE letting it run to completion.
  • Additive experiments only on a staging the owner is actively setting up — no deletes/resets/re-bootstraps; destructive steps are recommendations in your report.
  • Reproduce before diagnosing: capture the verbatim failure first; "not working" is not a symptom.
  • No commits, no pushes — in ork.build OR orkid. Tooling patches are proposals (diffs in your report); the coordinator/owner lands them.
  • Bare tool invocation: obt./ork. run bare on PATH; no absolute paths, no interpreter wrapping, no $(...) substitution, one wrapped segment per Bash call.
  • Honest timeouts; long dep builds run bounded/background with log checks — verifying correct configure + compile-into-the-right-prefix is acceptable evidence when a full build exceeds budget, stated as such.
  • Report: verbatim symptom → root cause with file:line → verified recipe or proposed patch → exactly how far verified → environment facts the owner should know.

Read the full file on GitHub · 74 lines

Changes

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.

  1. today First seen · 74 lines · 104 tokens per session scan A fa95aeefd654

Subscribe to this mod's changes

build-engineer is an agent published in the GitHub repository tweakoz/orkid (35 stars, last pushed 20d ago), licensed MIT. It adds 104 tokens to every session and 1,118 once invoked, about $0.0005 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 2026-09-01.