FE-1406, FE-1431, FE-1504, FE-1506, FE-1505, FE-1431, FE-1507, FE-1521, FE-1522, FE-1524: Consolidate the Brunch harness mission stack

[lunelson]

🌟 What is the purpose of this PR?

Land the accepted Brunch work from FE-1406 through FE-1524 through one stable review and merge surface. The ten-PR Graphite stack repeatedly lost approvals when a parent merged and GitHub rebased the remaining branches, so this PR preserves the sequence as ten ordered commits plus three follow-up fix commits on current main without making the consolidation branch part of another Graphite stack.

The final production throughline is the one established by the accepted missions: Petrinaut assistant panel → AI SDK /api/chat → Brunch-hosted plain Flue ChatAgent → model and tools → AI SDK stream → same Flue conversation, followed by an explicit harness-side capture sweep over a settled history range. The proof establishes that this integrated path is durable and that reapplying the range does not duplicate captures. It does not establish Mission 3's runbook/template/headless-PN path, extraction quality, a typed IR, session-as-net, or two brains.

The Brunch context root is libs/@hashintel/brunch-agent/. Its AGENTS.md defines the mission and PR-writing contract used below.

🔗 Related links

• Brunch context root: libs/@hashintel/brunch-agent/
• Accepted Mission 1 close report: docs/mission-archive/1-bare-petrinaut-flue-chat.md
• Accepted Mission 2 close report: docs/mission-archive/2-mechanical-capture-sweep.md
• Live, unimplemented-by-this-PR Mission 3 declaration: MISSION.md
• Original review history: #9327, #9337, #9345, #9346, #9347, #9348, #9349, #9381, #9382, #9384

🚫 Blocked by

• No external blocker; required CI and review are in progress.

🔍 What does this change?

• Preserves each original branch as one ordered commit while moving the complete stack onto current main.
• Establishes and then freezes the harness-owned plugin authoring surface and repertoire topology documented under the Brunch context root.
• Carries the Petrinaut panel integration, legible status, and per-purpose timing work through to the later topology.
• Replaces the active elicitation path with the accepted plain Petrinaut-to-Flue chat door and canonical Flue history.
• Adds the accepted mechanical capture-sweep pipe and proves idempotent reapplication on that live path.
• Archives accepted Missions 1 and 2 and cuts Mission 3 as the next execution authority without implementing FE-1525 here.
• Folds two current-toolchain test/lint corrections into the FE-1522 commit: the emitted-bundle assertion now recognizes Vite's named mount constant, and the history loader retains only the post-JSON cancellation guard.
• Carries stable transport idempotency keys into both Flue dispatch paths, so repeated initial and client-tool-result POSTs converge on their original submissions.
• Projects Flue's { awaiting: "client" } sentinel as a pending browser tool until a correlated client-tool-result signal supplies the actual output.
• Keeps the app and both shipped plugin workspaces in the pruned Brunch-core CI job because its architecture and contract tests inspect all three directly.
• Enables normal Turbo caching for all six Brunch build tasks; each already declares dist/** outputs, and no evidence-backed non-hermetic reason for the import-era opt-outs remains.

🏗️ Agent notes

Consolidation mission

Imperative: Get the accepted Brunch sequence through CI, review, and merge without another parent-merge rebase dismissing approvals. This is a delivery repair for the existing work, not a new product mission.

Throughline: current main → ten original branch deltas in order → one plain Git branch → one GitHub CI/review/merge surface. At the production boundary carried by those deltas: Petrinaut panel → transport-aisdk → Flue ChatAgent and canonical history → explicit harness apply-sweep → capture store keyed by Flue conversation identity → same range reapplied → same capture identities.

Proof: (1) the branch is based directly on current origin/main; (2) each original branch has one corresponding ordered commit; (3) the affected package tests, type checks, lint, builds, and architecture checks pass; and (4) the final Mission 1/2 throughline tests observe durable chat, tool return, history hydration, skill activation, capture creation, and idempotent reapplication. This proves a mergeable representation and the claims already closed by Missions 1 and 2. It does not prove Mission 3, improve extraction, join capture to an IR, or validate the approval workaround until GitHub actually merges this PR without rewriting it.

Constraints: Preserve semantic order and final content; keep FE-1525 out; keep this branch untracked by Graphite; retain the original PRs as review evidence until acceptance; preserve the topology gates in the Brunch context root; make no new harness mechanism for the sake of consolidation.

Fog-line: GitHub CI and merge behavior are still live evidence. Do not claim the approval-reset problem resolved until this plain branch passes required checks and merges without a history rewrite. Any failure must be discriminated as stack integration, current-main drift, or unrelated CI infrastructure before changing product code.

Stop or reorient: Stop if current main requires a semantic conflict resolution, if a correction changes an accepted mission's product behavior, if FE-1525 enters the branch, or if GitHub/Graphite begins rewriting this plain branch.

Observed proof and fog-line answers

• The ten branch deltas replayed onto current main without merge conflicts and remain ten commits in their original order, followed by three fix commits rather than a history rewrite.
• Before review fixes, the final tree differed from the original FE-1524 tip only by current-main changes and the two test/lint corrections named above. The additional commit now pins retry convergence and truthful pending-client-tool projection.
• All affected unit-test, type-check, and lint tasks pass locally; the Petrinaut build and architecture validation pass.
• The branch is confirmed untracked by Graphite.
• GitHub accepted the branch as one PR. Required CI and the eventual merge remain the unresolved part of the delivery proof.

Carried flags

• The original PRs remain open until this PR is accepted; then they can be closed as superseded.
• Mission 3 is now the context root's live MISSION.md, but FE-1525 and its implementation are intentionally excluded.
• This consolidation is a deliberate one-off exception to the context root's normal one-Linear-issue/one-Graphite-branch/one-PR lifecycle, forced by the broken approval/rebase integration. It should not become the ordinary Brunch workflow.

Included stack

Commit	Issue	Original PR	Scope
1	FE-1406	#9327	Establish the harness-teaching plugin contract
2	FE-1431	#9337	Build the plugin authoring surface
3	FE-1504	#9345	Run process-model elicitation from the Petrinaut panel
4	FE-1506	#9346	Show captures and completion during panel elicitation
5	FE-1505	#9347	Time interviewer turns by purpose
6	FE-1431	#9348	Complete and freeze the plugin authoring surface
7	FE-1507	#9349	Move the repertoire into core
8	FE-1521	#9381	Cut the agentic guidance layer to a mission charter
9	FE-1522	#9382	Replace elicitation with a Petrinaut-to-Flue chat
10	FE-1524	#9384	Prove idempotent capture sweep on the live Flue chat path

Pre-Merge Checklist 🚀

🚢 Has this modified a publishable library?

This PR:

• does not modify any publishable blocks or libraries, or modifications do not need publishing

The affected Brunch packages are private 0.0.0-private workspaces.

📜 Does this require a change to the docs?

The changes in this PR:

• require changes to docs which are made as part of this PR

User-facing guidance, ADRs, specifications, evidence, mission archives, and the context-root mission declaration are updated with the behavior they accompany.

🕸️ Does this require a change to the Turbo Graph?

The changes in this PR:

• affected the execution graph, and the turbo.json's have been updated to reflect this

⚠️ Known issues

• Running yarn workspace @apps/brunch-agent test:unit directly does not build workspace dependencies. Build @hashintel/brunch-agent-binding-flue first, or use the root Turbo test:unit task, whose dependency graph builds prerequisites.
• Required GitHub CI restarted for the build-cache change and is still in progress. No approval should be requested until it is green.
• History hydration now represents an unfinished browser tool truthfully as input-available, but AI SDK does not invoke onToolCall merely because an initial/hydrated message contains such a part. Full hard-reload execution and resume while readPetrinautDoc is pending remains unproven; this PR does not blanket-replay hydrated client tools because mutating tools would make that unsafe.

🐾 Next steps

• After this PR is accepted, close the ten original PRs as superseded while retaining them as review history.
• Continue Mission 3 on FE-1525 separately. Its runbook/template/headless-PN path must not be joined to Mission 2's capture store merely to tidy the architecture.
• Keep Host trunk, typed Petrinaut read/write, capture improvement, and any capture-to-IR convergence in MISSION.next.md until separately cut and re-earned.

🛡 What tests cover this?

• apps/brunch-agent/test/petrinaut-chat.test.ts drives the real Mission 1/2 entrypoint and covers the plain Flue chat door, server/client tools, canonical history, mounted skill, no interviewer sweep tool, capture creation, idempotent reapplication, duplicate initial/resume delivery convergence, and pending-client-tool history projection before resume.
• The Brunch core, binding-flue, transport-aisdk, SDCPN plugin, and Gherkin plugin unit suites cover the retained contract and topology layers.
• @apps/petrinaut-website unit tests cover principal/conversation identity, panel transport, and Flue-history hydration.
• apps/brunch-agent/test/flue-transcript.test.ts pins pending and completed browser-tool projection, including a completed null result.
• The emitted-artifact suite proves that the production bundle retains the agent registration, mounted route, database path, and client bundle.
• Petrinaut architecture validation checks 45 layers, 206 edges, 465 files, 46 generated pages, and 21 authored pages.
• .github/actions/prune-repository/prune_test.py pins the requested-scope extras, and a local reproduction of the CI prune followed by turbo run test:unit --env-mode=loose --filter @hashintel/brunch-agent passes all 18 files and 198 tests (197 passed, one skipped).
• Two identical filtered turbo run build invocations prove the cache change: the first run built all six Brunch tasks with cache misses; the second reported cache hits for all six (6 cached, 7 total, with the unrelated @hashintel/petrinaut-core build still opting out independently).

All test:unit, lint:eslint, and lint:tsc tasks pass for @hashintel/brunch-agent, @hashintel/brunch-agent-binding-flue, @hashintel/brunch-agent-transport-aisdk, @hashintel/brunch-agent-plugin-gherkin, @hashintel/brunch-agent-plugin-sdcpn, @apps/brunch-agent, and @apps/petrinaut-website. @hashintel/petrinaut also builds successfully.

❓ How to test this?

1. From the HASH root, run yarn install.
2. Run turbo run test:unit lint:eslint lint:tsc --filter=@hashintel/brunch-agent --filter=@hashintel/brunch-agent-binding-flue --filter=@hashintel/brunch-agent-transport-aisdk --filter=@hashintel/brunch-agent-plugin-gherkin --filter=@hashintel/brunch-agent-plugin-sdcpn --filter=@apps/brunch-agent --filter=@apps/petrinaut-website.
3. Run yarn workspace @local/petrinaut-arch-docs lint:arch-docs.
4. Inspect apps/brunch-agent/test/petrinaut-chat.test.ts: the same production door must show activate_skill, one capture envelope per named user entry, and the same capture ids with non-empty skippedDedupKeys on reapplication.
5. Optional human check: run the Brunch service and Petrinaut demo, send a panel message, reload to hydrate the same Flue conversation, and inspect the sibling <flueInstanceId>.json capture file. The interviewer must not expose a sweep tool.

📹 Demo

No new recording was made for this delivery-only consolidation. The durable proof is the production-path test and the accepted Mission 1 human-run record linked above.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
hash	Ready	Preview	Aug 31, 2026 11:09am
petrinaut	Ready	Preview	Aug 31, 2026 11:09am

2 Skipped Deployments

Project	Deployment	Actions	Updated (UTC)
hashdotdesign-tokens	Ignored	Preview	Aug 31, 2026 11:09am
petrinaut-docs	Skipped		Aug 31, 2026 11:09am

[cursor]

PR Summary

Medium Risk
Changes authentication and conversation ownership on agent routes plus the primary Petrinaut chat transport, though behavior is heavily covered by integration tests and explicit 401/403 gates.

Overview
This PR lands the accepted Brunch mission stack as one merge surface: the elicitation gallery (Gherkin/SDCPN elicitors, brunch_ask, target-document stores, and related integration tests) is removed in favor of a single plain Flue ChatAgent wired to Petrinaut’s /api/chat door.

Server and identity: Flue instance ids are SHA-256 hashes of x-brunch-principal plus conversation id; agentOwnershipGuard on /agents/chat/:id rejects missing identity (401) and hash mismatches (403). The chat handler now streams Flue chunks through a new AI SDK projector, supports GET history hydration, and resumes client tools via a client-tool-result signal (not elicitation ask/reply). Tools are ping, stub skill confirm-path, and browser-executed readPetrinautDoc (via @hashintel/petrinaut-core). Harness-side applyCaptureSweep writes JSON capture envelopes beside the sqlite DB; integration tests assert idempotent reapplication.

Petrinaut panel: The local-storage demo drops the brunch_ask interactive widget, pins stable principal and per-net conversation ids in localStorage, wraps the transport so useChat’s generated id does not replace the Flue conversation, and hydrates messages from GET /api/chat?id= before falling back to localStorage.

Docs, deps, CI: App dependencies drop direct elicitor plugins/core harness in favor of binding-flue + transport + petrinaut-core; OpenTelemetry is instrumented. .github/actions/prune-repository expands Brunch-core job fixtures (app, both plugins, evaluation paths). Brunch AGENTS.md is rewritten around mission contract (MISSION.md / archive). Large suites tied to the old path (walking skeleton, petrinaut-ask, duplicate transport-aisdk-server tests in the app) are deleted or replaced by expanded petrinaut-chat coverage.

^{Reviewed by Cursor Bugbot for commit 58058c7. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 2 potential issues.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 3704688. Configure here.}

[codspeed-hq]

Merging this PR will not alter performance

✅ 98 untouched benchmarks

---

_{Comparing ln/fe-1406-1524-consolidated (58058c7) with main (aae1779)}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 61.01%. Comparing base (3ebd68c) to head (58058c7).
⚠️ Report is 8 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #9463      +/-   ##
==========================================
- Coverage   61.01%   61.01%   -0.01%     
==========================================
  Files        1448     1449       +1     
  Lines      144648   144663      +15     
  Branches     6692     6692              
==========================================
  Hits        88262    88262              
- Misses      55282    55297      +15     
  Partials     1104     1104              

Flag	Coverage Δ
apps.hash-ai-worker-ts	1.99% <ø> (ø)
apps.hash-api	14.68% <ø> (ø)
blockprotocol.type-system	38.15% <ø> (-2.69%)	⬇️
local.claude-hooks	0.00% <ø> (ø)
local.harpc-client	51.49% <ø> (ø)
local.hash-backend-utils	3.27% <ø> (ø)
local.hash-graph-sdk	10.02% <ø> (ø)
local.hash-isomorphic-utils	12.22% <ø> (ø)
rust.antsi	2.36% <ø> (ø)
rust.error-stack	90.81% <ø> (ø)
rust.harpc-codec	84.70% <ø> (ø)
rust.harpc-net	96.19% <ø> (ø)
rust.harpc-tower	67.03% <ø> (ø)
rust.harpc-types	0.00% <ø> (ø)
rust.harpc-wire-protocol	92.23% <ø> (ø)
rust.hash-codec	72.76% <ø> (ø)
rust.hash-config	74.74% <ø> (ø)
rust.hash-graph-api	19.76% <ø> (ø)
rust.hash-graph-authentication	97.00% <ø> (ø)
rust.hash-graph-authorization	63.14% <ø> (ø)
rust.hash-graph-embeddings	91.88% <ø> (ø)
rust.hash-graph-postgres-store	32.02% <ø> (ø)
rust.hash-graph-store	48.41% <ø> (ø)
rust.hash-graph-temporal-versioning	50.18% <ø> (ø)
rust.hash-graph-types	0.00% <ø> (ø)
rust.hash-graph-validation	84.71% <ø> (ø)
rust.hash-middleware	92.19% <ø> (ø)
rust.hashql-ast	89.63% <ø> (ø)
rust.hashql-compiletest	28.39% <ø> (ø)
rust.hashql-core	78.95% <ø> (ø)
rust.hashql-diagnostics	72.51% <ø> (ø)
rust.hashql-eval	79.82% <ø> (ø)
rust.hashql-hir	89.09% <ø> (ø)
rust.hashql-mir	87.92% <ø> (ø)
rust.hashql-syntax-jexpr	94.04% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[lunelson]

Enabled normal Turbo caching for all six Brunch build tasks in 58058c71a7. The opt-outs arrived wholesale during the FE-1437 monorepo import and had no recorded non-hermetic justification. Verification used two identical filtered builds: first run missed all six Brunch caches; second run hit all six (6 cached, 7 total, with only unrelated @hashintel/petrinaut-core still bypassing its own cache).

[TimDiekmann]

The failing Unit (@hashintel/brunch-agent) job was a pruned-checkout mismatch, not a product-test regression: core tests read the shipped SDCPN/Gherkin contracts and topology, while the prune rule retained only core and the app. Fixed in d2ded94308 by adding both plugin workspaces only when Brunch core is the requested scope. The hermetic prune-rule test passes, and an exact local prune followed by turbo run test:unit --env-mode=loose --filter @hashintel/brunch-agent now passes 18 test files / 198 tests (197 passed, one skipped).

Note, that on main we have merged the change in Turborepo to use affectedTasks instead of affectedPackages. So if you don't strictly need a dependency, you may just add a dependsOn relationship in turbo.json.

[github-actions]

Benchmark results

@rust/hash-graph-benches – Integrations

policy_resolution_large

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 2002	$$16.3 \mathrm{ms} \pm 238 \mathrm{μs}\left({\color{lightgreen}-5.327 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$2.08 \mathrm{ms} \pm 22.7 \mathrm{μs}\left({\color{gray}0.246 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 1002	$$7.71 \mathrm{ms} \pm 61.3 \mathrm{μs}\left({\color{lightgreen}-8.347 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 3314	$$27.8 \mathrm{ms} \pm 256 \mathrm{μs}\left({\color{gray}3.47 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$9.40 \mathrm{ms} \pm 98.6 \mathrm{μs}\left({\color{gray}1.20 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 1527	$$15.3 \mathrm{ms} \pm 109 \mathrm{μs}\left({\color{gray}-2.187 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 2078	$$17.5 \mathrm{ms} \pm 137 \mathrm{μs}\left({\color{gray}-4.454 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$2.34 \mathrm{ms} \pm 18.8 \mathrm{μs}\left({\color{gray}0.465 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 1033	$$9.74 \mathrm{ms} \pm 81.8 \mathrm{μs}\left({\color{gray}0.726 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_medium

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 102	$$2.45 \mathrm{ms} \pm 20.3 \mathrm{μs}\left({\color{red}6.38 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$1.98 \mathrm{ms} \pm 18.1 \mathrm{μs}\left({\color{gray}3.87 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 52	$$2.27 \mathrm{ms} \pm 35.4 \mathrm{μs}\left({\color{red}6.12 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 269	$$3.12 \mathrm{ms} \pm 21.1 \mathrm{μs}\left({\color{lightgreen}-5.414 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$2.14 \mathrm{ms} \pm 23.6 \mathrm{μs}\left({\color{gray}1.06 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 108	$$2.53 \mathrm{ms} \pm 20.1 \mathrm{μs}\left({\color{gray}-2.024 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 133	$$2.82 \mathrm{ms} \pm 30.9 \mathrm{μs}\left({\color{red}7.73 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$2.23 \mathrm{ms} \pm 15.8 \mathrm{μs}\left({\color{red}7.03 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 63	$$2.61 \mathrm{ms} \pm 19.4 \mathrm{μs}\left({\color{red}5.57 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_none

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 2	$$1.60 \mathrm{ms} \pm 13.4 \mathrm{μs}\left({\color{gray}-4.953 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$1.65 \mathrm{ms} \pm 27.4 \mathrm{μs}\left({\color{red}5.71 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 2	$$1.61 \mathrm{ms} \pm 15.1 \mathrm{μs}\left({\color{gray}-2.240 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 8	$$1.81 \mathrm{ms} \pm 23.5 \mathrm{μs}\left({\color{gray}-3.137 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$1.64 \mathrm{ms} \pm 12.5 \mathrm{μs}\left({\color{gray}-4.688 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 3	$$1.76 \mathrm{ms} \pm 15.1 \mathrm{μs}\left({\color{gray}-4.803 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_small

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 52	$$1.94 \mathrm{ms} \pm 26.7 \mathrm{μs}\left({\color{red}5.02 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$1.77 \mathrm{ms} \pm 13.2 \mathrm{μs}\left({\color{gray}1.41 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 26	$$1.91 \mathrm{ms} \pm 22.5 \mathrm{μs}\left({\color{red}5.87 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 94	$$2.17 \mathrm{ms} \pm 23.7 \mathrm{μs}\left({\color{gray}1.99 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$1.83 \mathrm{ms} \pm 11.5 \mathrm{μs}\left({\color{gray}-4.074 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 27	$$2.00 \mathrm{ms} \pm 12.5 \mathrm{μs}\left({\color{gray}-1.130 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 66	$$2.08 \mathrm{ms} \pm 22.6 \mathrm{μs}\left({\color{gray}3.88 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$1.89 \mathrm{ms} \pm 15.8 \mathrm{μs}\left({\color{gray}4.16 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 29	$$2.03 \mathrm{ms} \pm 19.4 \mathrm{μs}\left({\color{gray}3.70 \mathrm{\%}}\right) $$	Flame Graph

read_scaling_complete

Function	Value	Mean	Flame graphs
entity_by_id;one_depth	1 entities	$$26.0 \mathrm{ms} \pm 275 \mathrm{μs}\left({\color{gray}-0.105 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	10 entities	$$22.3 \mathrm{ms} \pm 145 \mathrm{μs}\left({\color{red}5.35 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	25 entities	$$23.3 \mathrm{ms} \pm 174 \mathrm{μs}\left({\color{gray}-2.696 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	5 entities	$$21.0 \mathrm{ms} \pm 187 \mathrm{μs}\left({\color{gray}1.94 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	50 entities	$$26.6 \mathrm{ms} \pm 156 \mathrm{μs}\left({\color{gray}-4.665 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	1 entities	$$30.7 \mathrm{ms} \pm 216 \mathrm{μs}\left({\color{gray}1.81 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	10 entities	$$26.2 \mathrm{ms} \pm 213 \mathrm{μs}\left({\color{red}6.40 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	25 entities	$$59.2 \mathrm{ms} \pm 629 \mathrm{μs}\left({\color{red}6.25 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	5 entities	$$21.7 \mathrm{ms} \pm 204 \mathrm{μs}\left({\color{gray}1.57 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	50 entities	$$151 \mathrm{ms} \pm 922 \mathrm{μs}\left({\color{gray}4.36 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	1 entities	$$6.70 \mathrm{ms} \pm 43.2 \mathrm{μs}\left({\color{gray}0.098 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	10 entities	$$6.94 \mathrm{ms} \pm 77.9 \mathrm{μs}\left({\color{red}5.61 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	25 entities	$$7.16 \mathrm{ms} \pm 62.0 \mathrm{μs}\left({\color{gray}3.85 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	5 entities	$$6.81 \mathrm{ms} \pm 79.1 \mathrm{μs}\left({\color{red}5.15 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	50 entities	$$7.15 \mathrm{ms} \pm 52.7 \mathrm{μs}\left({\color{red}7.90 \mathrm{\%}}\right) $$	Flame Graph

read_scaling_linkless

Function	Value	Mean	Flame graphs
entity_by_id	1 entities	$$6.97 \mathrm{ms} \pm 77.4 \mathrm{μs}\left({\color{red}6.51 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	10 entities	$$6.95 \mathrm{ms} \pm 113 \mathrm{μs}\left({\color{gray}4.55 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	100 entities	$$7.45 \mathrm{ms} \pm 66.8 \mathrm{μs}\left({\color{red}10.9 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	1000 entities	$$7.49 \mathrm{ms} \pm 82.7 \mathrm{μs}\left({\color{red}9.10 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	10000 entities	$$7.58 \mathrm{ms} \pm 69.2 \mathrm{μs}\left({\color{red}10.6 \mathrm{\%}}\right) $$	Flame Graph

representative_read_entity

Function	Value	Mean	Flame graphs
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/block/v/1	$$7.06 \mathrm{ms} \pm 82.0 \mathrm{μs}\left({\color{gray}-1.035 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/book/v/1	$$7.99 \mathrm{ms} \pm 57.3 \mathrm{μs}\left({\color{red}15.0 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/building/v/1	$$7.86 \mathrm{ms} \pm 564 \mathrm{μs}\left({\color{red}16.5 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/organization/v/1	$$7.51 \mathrm{ms} \pm 93.5 \mathrm{μs}\left({\color{red}10.1 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/page/v/2	$$7.56 \mathrm{ms} \pm 77.5 \mathrm{μs}\left({\color{red}13.1 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/person/v/1	$$8.01 \mathrm{ms} \pm 71.6 \mathrm{μs}\left({\color{red}17.6 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/playlist/v/1	$$7.01 \mathrm{ms} \pm 70.8 \mathrm{μs}\left({\color{red}5.83 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/song/v/1	$$7.31 \mathrm{ms} \pm 45.4 \mathrm{μs}\left({\color{red}8.30 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/uk-address/v/1	$$7.58 \mathrm{ms} \pm 68.8 \mathrm{μs}\left({\color{red}13.1 \mathrm{\%}}\right) $$	Flame Graph

representative_read_entity_type

Function	Value	Mean	Flame graphs
get_entity_type_by_id	Account ID: bf5a9ef5-dc3b-43cf-a291-6210c0321eba	$$5.63 \mathrm{ms} \pm 38.1 \mathrm{μs}\left({\color{red}11.4 \mathrm{\%}}\right) $$	Flame Graph

representative_read_multiple_entities

Function	Value	Mean	Flame graphs
entity_by_property	traversal_paths=0	0	$$41.7 \mathrm{ms} \pm 269 \mathrm{μs}\left({\color{red}5.41 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=255	1,resolve_depths=inherit:1;values:255;properties:255;links:127;link_dests:126;type:true	$$75.6 \mathrm{ms} \pm 971 \mathrm{μs}\left({\color{gray}3.68 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:0;link_dests:0;type:false	$$47.6 \mathrm{ms} \pm 598 \mathrm{μs}\left({\color{red}7.10 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:1;link_dests:0;type:true	$$51.9 \mathrm{ms} \pm 440 \mathrm{μs}\left({\color{gray}1.46 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:2;links:1;link_dests:0;type:true	$$61.2 \mathrm{ms} \pm 507 \mathrm{μs}\left({\color{red}7.81 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:2;properties:2;links:1;link_dests:0;type:true	$$62.6 \mathrm{ms} \pm 478 \mathrm{μs}\left({\color{gray}1.65 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=0	0	$$29.2 \mathrm{ms} \pm 889 \mathrm{μs}\left({\color{gray}1.93 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=255	1,resolve_depths=inherit:1;values:255;properties:255;links:127;link_dests:126;type:true	$$44.8 \mathrm{ms} \pm 332 \mathrm{μs}\left({\color{lightgreen}-9.185 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:0;link_dests:0;type:false	$$32.3 \mathrm{ms} \pm 322 \mathrm{μs}\left({\color{gray}-3.849 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:1;link_dests:0;type:true	$$38.5 \mathrm{ms} \pm 271 \mathrm{μs}\left({\color{gray}-1.960 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:2;links:1;link_dests:0;type:true	$$39.0 \mathrm{ms} \pm 483 \mathrm{μs}\left({\color{lightgreen}-7.465 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:2;properties:2;links:1;link_dests:0;type:true	$$37.6 \mathrm{ms} \pm 247 \mathrm{μs}\left({\color{lightgreen}-8.867 \mathrm{\%}}\right) $$

scenarios

Function	Value	Mean	Flame graphs
full_test	query-limited	$$71.6 \mathrm{ms} \pm 345 \mathrm{μs}\left({\color{lightgreen}-5.612 \mathrm{\%}}\right) $$	Flame Graph
full_test	query-unlimited	$$79.1 \mathrm{ms} \pm 341 \mathrm{μs}\left({\color{gray}-2.670 \mathrm{\%}}\right) $$	Flame Graph
linked_queries	query-limited	$$13.7 \mathrm{ms} \pm 130 \mathrm{μs}\left({\color{red}25.3 \mathrm{\%}}\right) $$	Flame Graph
linked_queries	query-unlimited	$$327 \mathrm{ms} \pm 1.03 \mathrm{ms}\left({\color{gray}3.82 \mathrm{\%}}\right) $$	Flame Graph