runtime-profile: publish default bootstrap resources by idy · Pull Request #9 · GizClaw/raids

idy · 2026-07-24T01:43:29Z

Summary

publish the canonical, applyable RuntimeProfile/default using the current GizClaw Desktop default composition
publish RegistrationToken/default-runtime with public client token default
validate catalog identities, reference closure, selected Workflow aliases, duplicate tokens, placeholders, and prohibited deployment fields
document dependency ordering, per-Server resource ownership, and the documentation-only example boundary

Validation

python scripts/validate_catalog.py
python -m unittest discover -s tests -v
ruff format --check scripts tests
ruff check scripts tests
git diff --check
GizClaw main@95d08074 isolated e2e:
- applying the token before the profile fails with an unresolved profile error
- first profile/token apply returns created
- second profile/token apply returns unchanged
- a peer registering with token default receives RuntimeProfile/default

Closes #7

Raids needs one canonical public runtime composition and stable bootstrap selector so consumers do not independently recreate the same contract. - Publish RuntimeProfile/default with the complete public resource closure - Add RegistrationToken/default-runtime with the public token value default - Validate catalog identities, references, workflow aliases, and prohibited defaults - Document dependency order, per-Server ownership, and the example boundary Generated with [Codex](https://github.com/openai)

github-actions

🤖 OpenAI PR review

Reviewed commit: 2ff594129c

Reviewed range: d52656c3d7..2ff594129c

Review mode: full

Session key: repo-1308402561-pr-9-v2

Generation key: b8d95a3a9b4eaaadde27123bebe634b695f5bafb8cb69ab5fbe4647347a29da1

Diff chunks: 1

Model: gpt-5.6-terra

Reasoning effort: medium

Review time: 50s

Input tokens: 157,071

Cached input tokens: 131,468

Cache write tokens: 25,579

Cache hit ratio: 83.7%

Output tokens: 4,314

Reasoning tokens: 3,098

Total tokens: 161,385

Estimated Codex credits: 4.040

Credit rate per 1M tokens: 62.5 input / 6.25 cached / 375 output

Per-chunk and aggregation usage

Turn key	Time	Input	Cached	Cache hit	Cache write	Output	Reasoning	Total	Est. credits
`chunk:1/1:2f55cae5e61e4d78a51cd4d979ca47a97cf6d55d14e91e2f181628521bf35882`	46s	106,504	82,806	77.7%	23,680	4,011	3,063	110,515	3.503
`aggregate:b8d95a3a9b4eaaadde27123bebe634b695f5bafb8cb69ab5fbe4647347a29da1`	4s	50,567	48,662	96.2%	1,899	303	35	50,870	0.537

Found 1 actionable validation gap.

1 inline finding published.

Registration tokens are independently publishable catalog resources, so each one must resolve its target profile before consumers apply it. - Check every token runtime_profile_name against collected RuntimeProfiles - Reject missing or empty profile selectors - Add a regression test for a non-default token with a missing profile Generated with [Codex](https://github.com/openai)

idy · 2026-07-24T01:48:08Z

@codex review

github-actions

🤖 OpenAI PR review

Reviewed commit: 9fa7df69e9

Reviewed range: 2ff594129c..9fa7df69e9

Review mode: incremental

Session key: repo-1308402561-pr-9-v2

Generation key: 22a7274d4169f5a86596ec6080614a7bcadd30bb9985e1429c12c630d8feb742

Diff chunks: 1

Model: gpt-5.6-terra

Reasoning effort: medium

Review time: 13s

Input tokens: 109,109

Cached input tokens: 81,023

Cache write tokens: 28,074

Cache hit ratio: 74.3%

Output tokens: 542

Reasoning tokens: 232

Total tokens: 109,651

Estimated Codex credits: 2.465

Credit rate per 1M tokens: 62.5 input / 6.25 cached / 375 output

Per-chunk and aggregation usage

Turn key	Time	Input	Cached	Cache hit	Cache write	Output	Reasoning	Total	Est. credits
`chunk:1/1:8f4b6abb163ca5199b41924cf7e6b215e02f6e544b4de6f325c779e4c71188ea`	9s	53,371	26,111	48.9%	27,254	391	222	53,762	2.014
`aggregate:22a7274d4169f5a86596ec6080614a7bcadd30bb9985e1429c12c630d8feb742`	4s	55,738	54,912	98.5%	820	151	10	55,889	0.451

No actionable findings.

idy

The default RuntimeProfile and bootstrap token composition looks coherent, the existing profile-reference finding is fixed, and the advertised catalog/unit/Ruff checks pass locally. I found one remaining actionable validation gap on arbitrary RegistrationToken resources and left it inline.

idy · 2026-07-24T02:10:57Z

Reviewed the latest head 9fa7df6. I found one blocking catalog-validation gap. The published RuntimeProfile/default itself is byte-for-byte identical to GizClaw origin/main's current Desktop default profile, and the catalog validator, 7 unit tests, Ruff 0.14.14, git diff --check, and all current PR checks pass.

[P1] Normalize token values before duplicate detection — scripts/validate_catalog.py:342

The duplicate map is keyed by the raw YAML string, but GizClaw normalizes RegistrationToken.token with strings.TrimSpace before hashing and enforcing uniqueness. As a result, two resources with token: other and token: " other " both pass this validator, while applying the second one to a clean Server returns a 409 because both normalize to other. That violates #7's acceptance criterion that repository validation reject duplicate token values and permits a release that passes CI but cannot be fully applied. Normalize the token exactly as the Server does (also rejecting an empty normalized value) before checking token_values, and add a regression test for a whitespace-only collision.

Catalog validation must use the same token normalization and required-field semantics as GizClaw Server or invalid manifests can pass CI and fail during apply. - Normalize token values with the Go TrimSpace character set before duplicate detection - Reject missing, empty, and non-string token values - Cover ASCII and Unicode whitespace collisions plus invalid token shapes Generated with [Codex](https://github.com/openai)

idy · 2026-07-24T02:21:03Z

Addressed the additional token-normalization finding in da222cc: catalog validation now mirrors Go strings.TrimSpace before token duplicate detection and rejects missing, empty, or non-string token values. Added regression coverage for ASCII and Unicode whitespace collisions plus invalid token shapes. Verified all 1,033 YAML files, 9 unit tests, Ruff 0.14.14, and git diff --check.

idy · 2026-07-24T02:21:20Z

@codex

github-actions

🤖 OpenAI PR review

Reviewed commit: da222cccdc

Reviewed range: 9fa7df69e9..da222cccdc

Review mode: incremental

Session key: repo-1308402561-pr-9-v2

Generation key: 077d46a6ff92bc5a2ebd0e91fbba77854c819bc80202b953f204d225bbb49dc4

Diff chunks: 1

Model: gpt-5.6-terra

Reasoning effort: medium

Review time: 11s

Input tokens: 120,548

Cached input tokens: 89,438

Cache write tokens: 31,098

Cache hit ratio: 74.2%

Output tokens: 534

Reasoning tokens: 224

Total tokens: 121,082

Estimated Codex credits: 2.704

Credit rate per 1M tokens: 62.5 input / 6.25 cached / 375 output

Per-chunk and aggregation usage

Turn key	Time	Input	Cached	Cache hit	Cache write	Output	Reasoning	Total	Est. credits
`chunk:1/1:8fadd324944a4e9bb57a5657c8d0cc55bc68668ccbf3336b018331bdbb35387e`	7s	58,749	28,451	48.4%	30,292	386	217	59,135	2.216
`aggregate:077d46a6ff92bc5a2ebd0e91fbba77854c819bc80202b953f204d225bbb49dc4`	4s	61,799	60,987	98.7%	806	148	7	61,947	0.487

No actionable findings.

idy · 2026-07-24T02:48:15Z

Re-reviewed the complete PR at da222cc. No new actionable or blocking findings remain.

The follow-up now mirrors Go strings.TrimSpace semantics before token uniqueness checks, keys duplicate detection by the normalized value, rejects missing/empty/non-string tokens, and adds regression coverage for ASCII and Unicode whitespace collisions. I independently reran the catalog validator over all 1,033 YAML files, all 9 unit tests, Ruff 0.14.14, and git diff --check; everything passed. The published RuntimeProfile/default remains byte-for-byte identical to GizClaw origin/main's current Desktop default profile, all PR checks are green, and the PR title/three review-ready commits remain coherent.

The clean-server apply/register E2E was not rerun in this pass; this review relies on the successful E2E recorded in the PR description for that path. From the Raids catalog and current Desktop-consumer contract perspective, this is ready to merge.

github-actions Bot reviewed Jul 24, 2026

View reviewed changes

Comment thread scripts/validate_catalog.py

github-actions Bot reviewed Jul 24, 2026

View reviewed changes

idy commented Jul 24, 2026

View reviewed changes

Comment thread scripts/validate_catalog.py Outdated

github-actions Bot reviewed Jul 24, 2026

View reviewed changes

idy merged commit 2981f6b into main Jul 24, 2026
3 checks passed

idy deleted the codex/issue-7-default-runtime-bootstrap branch July 24, 2026 03:23

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

runtime-profile: publish default bootstrap resources#9

runtime-profile: publish default bootstrap resources#9
idy merged 3 commits into
mainfrom
codex/issue-7-default-runtime-bootstrap

idy commented Jul 24, 2026

Uh oh!

github-actions Bot left a comment

Uh oh!

Uh oh!

idy commented Jul 24, 2026

Uh oh!

github-actions Bot left a comment

Uh oh!

idy left a comment

Uh oh!

Uh oh!

idy commented Jul 24, 2026

Uh oh!

idy commented Jul 24, 2026

Uh oh!

idy commented Jul 24, 2026

Uh oh!

github-actions Bot left a comment

Uh oh!

idy commented Jul 24, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Conversation

idy commented Jul 24, 2026

Summary

Validation

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

🤖 OpenAI PR review

Uh oh!

Uh oh!

idy commented Jul 24, 2026

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

🤖 OpenAI PR review

Uh oh!

idy left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

idy commented Jul 24, 2026

Uh oh!

idy commented Jul 24, 2026

Uh oh!

idy commented Jul 24, 2026

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

🤖 OpenAI PR review

Uh oh!

idy commented Jul 24, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant