Loops, autoprompt and waves

01

Governed loop

An HPP loop has:

1. objective and spec;
2. observable state;
3. next action;
4. time/iteration budget;
5. evidence gate;
6. stop conditions;
7. human escalation;
8. persisted feedback.

loop_gate.py implements the stop gate: a textual completion mark does not release the loop unless the corresponding done_gate passes. autoprompt_resume.py produces resumption; it does not change the verdict.

For an end-to-end or visual criterion the latch is hpp evidence run (new in 2.6.0): it re-executes the declared command, measures its exit code outside the model and hashes the artifacts it left. hpp evidence verify is later reconciliation — it re-derives a record from the files on disk — and never the latch: the record's self-hash makes an edit visible and is not a signature. hpp evidence run --sign-key KEY --signer NAME adds an SSH signature, and hpp evidence verify --allowed-signers FILE refuses a record that does not carry one the file accepts; a signature says who wrote the record, not that the command still passes, so whoever must not trust the maker runs the command again. hpp drives no browser; the browser runs inside the command you declare.

02

Autoloop and LoopGraph

Autoloop is the finite cycle observe → choose → act → verify → record → stop/continue. In HPP it does not mean autonomy without a ceiling. The loop in hpp.manifest.json is the minimal LoopGraph: states, events, gates and allowed transitions. Intentional repetition lives in the charter with a budget and an exit; an accidental cycle in the WorkGraph is an error.

03

Spec-driven

The spec is compiled into a WorkGraph. Acceptance criteria travel with each unit. Dependencies become topological waves and each wave closes tests and review before releasing the next one.

04

Autoprompt

Autoprompt answers "what is the next step derivable from the state?". It does not decide risk, does not approve a sensitive change and does not replace a checker. If the state is insufficient, the correct output is to declare the missing data.

05

Failure memory

Gotcha Memory classifies failures, measures recurrence and proposes a lesson. Promotion is controlled to avoid turning an isolated or ambiguous incident into a permanent rule.

A failure that matches no family stays unknown and is never relabelled automatically. An advisory relabel through hpp decide (new in 2.6.0) is offline: a person runs it, never a hook, after measuring the decider with hpp decide eval; the record stays advisory, and a new family enters the classifier only as a reviewed change.

06

Evaluation

pass@k measures whether the system can; pass^k measures whether it repeats. The loop only treats release-critical work as stable when the regression gate passes in the declared universe and k.

The same rule applies to the instruments a loop leans on. hpp retrieval eval measures a retriever and hpp decide eval a decider (both new in 2.6.0), each on labelled cases, with instrument failures — a timeout, a crash, a malformed answer — counted apart and never scored. Best-of-N, where N lanes build alternatives and a reviewer from another lane and model family selects one (lane_board.py compete and select, new in 2.6.0), is pass@N: choosing 1 of N measures that the system can, not that it repeats, so the winner still needs its own verification and pass^k.