feat(elision): invariant-bearing truncation markers; bump tok to #82 - #235
Merged
Conversation
Bumps the tok submodule to a1d1863f (tok#82: verified-fact elision
summaries) and wires it into every tool-output truncation site.
Hawk's markers previously said only '... (truncated)'. Measured agent
behavior shows bare-count markers cause 11-97 retrieval-call storms,
because the reader cannot tell whether the answer was in the dropped
portion. Every marker now carries facts VERIFIED across the elided
units and nothing else:
- JSON tool output: '{sku-114..159} N records elided: status=shipped
x46, range id=...' — field constants, enumerations summing to the
total, numeric ranges, distinct-count coverage, dense-run upgrades.
Structural cuts now splice at TOP-LEVEL commas (depth-aware scan)
so kept records stay complete and the dropped tail parses as whole
records; mid-record raw cuts still fail closed to the bare marker.
- Log-shaped output: level distribution ('N lines elided: info x6').
- Prose/unknown: unchanged bare marker.
Fail-closed preserved throughout: anything unparseable or under the
3-unit threshold truncates exactly as before. New elision.go exposes
the shared notice builder; internal/token facades the two new tok
primitives. Full engine suite green.
The release-parity check resolves each submodule's go.mod version and requires it to equal the index gitlink. Bumping the gitlink without updating the module requirement left CI resolving old tok (643b6675), which lacks JSONInvariants/LogInvariants — breaking module hygiene builds. Pin the require directive to a1d1863f (#82).
Patel230
added a commit
that referenced
this pull request
Aug 23, 2026
Rebase onto post-#235 main; update the tok require directive to the toolschema merge and align the gitlink, satisfying release parity.
Patel230
added a commit
that referenced
this pull request
Aug 23, 2026
#236) * feat(tools): opt-in tool-catalog shrink via tok; bump submodule to #83 Wires tok's new toolschema compressor (GrayCodeAI/tok#83) into the outgoing request path behind HAWK_TOOL_SHRINK=1 (default off, request bytes unchanged). At ChatService.BuildOptions the tool list converts to the OpenAI function-catalog wire shape, shrinks, and converts back. Fail-open guarantees: any marshal/unmarshal error, tool-count drift, or name mismatch at any position returns the original list untouched; disabled or not-smaller catalogs are no-ops. When compression lands, the exact pre-shrink catalog persists under state/tool-catalog-originals/<sha>.json keyed by content hash for recovery and diffing, and a one-time info log reports before/after byte counts. Facades extended: internal/token aliases tok.ToolShrinkStats and exposes ShrinkToolCatalog/LintToolCatalog; internal/engine/token re-exports. Submodule bump includes tok#82 invariants (already wired separately on feat/invariant-markers; this branch stacks on it for the pointer). New tests: disabled-by-default no-op, enabled reduction with name/ constraint/required preservation through BuildOptions, and recovery-file persistence. * fix(deps): pin tok module to #83 (a7c4b99d) for submodule parity Rebase onto post-#235 main; update the tok require directive to the toolschema merge and align the gitlink, satisfying release parity.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Phase 1 of the caveman-style compression adoption (approved plan): every tool-output truncation marker now states only verified facts about the elided portion, replacing bare
"… (truncated)". Measured agent behavior shows bare-count markers cause 11–97 retrieval-call storms because readers can't tell whether the answer was dropped.Prerequisite merged: GrayCodeAI/tok#82 (
invariants.go— fact computation + sentinel upgrades). This PR bumps the submodule pointer and wires hawk to it.Changes
a1d1863f:JSONInvariants/LogInvariantsprimitives;CompressJSONsentinel record; enriched log markers.internal/engine/elision.go: shared notice builder — JSON arrays get field-level facts (constants, enumerations summing to total, numeric ranges, distinct-count coverage, dense-run upgrades); log-shaped text gets level distribution; prose falls back unchanged.truncateToolOutput, spill preview tail, and all 5 cut paths intruncateOutputStructurally.internal/token+internal/engine/tokenexpose the two new primitives.Testing
New elision suite (6 tests: JSON facts, log distribution, fragment normalization incl. trailing-bracket/comma artifacts, fail-closed fallbacks) + full engine suite green; lint/vet clean.