Agenticcoding

One Claude Code MCP call fans out to paid workers while a shared cost capability blocks work beyond the 25 dollar limit

One Claude Code MCP call spent the whole agent budget

The Claude Code run had a $25 limit and made one MCP call. Hidden fan-out behind the server produced a $286.40 bill.

A Claude Code MCP write waits in a retry queue beyond its approval expiry, keeps the same operation ID, and is blocked by an execution-time authority check

Claude Code queued an approved MCP write. It ran after permission expired

A stable operation ID can stop duplicate MCP writes while still allowing an action after its approval expires. Recheck authority where the effect occurs.

A Claude Code MCP write passes through an acknowledgement gate, durable commit check, and independent read before a durability receipt is issued

Claude Code got a success receipt before the write was durable

An MCP tool can return success after accepting a write but before durable storage. Define acknowledgement levels, verify the committed version, and give Claude Code a durability receipt.

Two parallel Claude Code writes request operation IDs from one atomic identity ledger, which issues unique IDs and blocks collisions before MCP dispatch

Claude Code gave two parallel writes the same operation ID

Two parallel Claude Code tool calls can share an operation ID and corrupt retry evidence. Allocate identity centrally, bind it to intent, and reject collisions before dispatch.

Two Claude Code eval cases are separated by a reset gate that verifies a clean sandbox before the second case starts

Claude Code passed the eval because the previous case cleaned up for it

A Claude Code eval can pass only because an earlier case left the sandbox in a helpful state. Give every case a fresh environment, a state fingerprint, and a reset receipt.

Three Claude Code subagents request reservations from one parent-owned budget ledger, which admits two calls and blocks the third before the shared ceiling is exceeded

Claude Code stayed under every child budget and still overspent

Three Claude Code subagents each stayed inside their local limit. Together they broke the run budget because nobody reserved spend at the parent boundary.

Claude Code course path from a convincing demo to a governed run and production evidence

My Claude Code production agents course is live

My new practical Leanpub course turns the production engineering behind Claude Code agents into labs, assessments, capstones, and a production-readiness dossier.

A Claude Code patch enters an independent evidence review before the agent explanation is revealed in a second pass

Do not let Claude Code explain its patch to the evaluator first

A polished Claude Code handoff can anchor the reviewer before the evidence is checked. Give an independent evaluator the task contract, patch, and raw test results first, then reveal the agent’s explanation.

A Claude Code trace with a missing MCP tool event is blocked by a completeness gate before replay and review

Make Claude Code prove its trace is complete before replay

A clean replay can be dangerously reassuring when the original trace dropped the tool call that changed production. Add sequence checks, expected-event accounting, and a completeness receipt before review.

Claude Code compacts its conversation while an external control-state checkpoint preserves authority, denied actions, pending effects, and evidence bindings

Keep Claude Code's safety state outside its context window

Context compaction can preserve the task while losing the decision that blocked a dangerous tool call. Store authority, denials, and unresolved effects in a control checkpoint outside the transcript.