[uk-ai-advisory:operational-resilience] UK AI advisory: operational resilience for githubnext/gh-cao
メンテナーはふだん 1 日以内に返信
まだ誰も着手していません。
評価
- 難易度
- 2/5
- 見積もり時間
- 1〜3時間
- 初心者へのやさしさ
- 68/100
調査の方向性
ルートのgo.modにあるgithub.com/klauspost/compressのエントリから始め、v1.18.7以降に更新してください。次に、ルートでGoのビルドとテストを実行します。ビルドとテストが成功し、Dependabot alert 103が未解決でなくなれば、issueは完了です。issueでは作業をCopilotに割り当てることが推奨されていますが、担当者もリンクされたプルリクエストも記載されていません。
索引モデルが issue の本文から書いたものです。
説明
githubnext/gh-cao is a public, MIT-licensed, actively maintained repository. This advisory proposes it as a B (Open With Conditions) candidate and recommends keeping the code open. Nine high-severity alerts are open across Dependabot and CodeQL. The most valuable fix is one dependency bump: github.com/klauspost/compress is pinned at v1.18.0 in the root go.mod, and v1.18.7 is the first patched version for Dependabot alert 103. Six CodeQL cache-poisoning alerts on release-promote.yml have been open for 25 days, and no patch SLA is evidenced to judge them against.
This is a non-binding, seven-day advisory with explicit limits. It cannot observe organizational controls, and it is not an authorization, a security assessment, or a recommendation to close code.
Action: A maintainer should assign this issue to Copilot using the Agent prompt under Recommended Next Action. Acceptance check: the root go.mod requires github.com/klauspost/compress v1.18.7 or later, the root Go build and tests pass, and Dependabot alert 103 is no longer open.
Advisory Status
HUMAN_REVIEW_REQUIRED
Critical findings:
- Dependabot alert 103 (high): the root
go.modpinsgithub.com/klauspost/compressv1.18.0, which is inside the vulnerable range (1.18.6 and earlier). The first patched version is 1.18.7. The alert was opened less than a day before this run. - CodeQL alerts 39–44 (high, 25 days):
actions/cache-poisoning/poisonable-stepin.github/workflows/release-promote.yml. The workflow comments that cache mode is read-only, but no dismissal or verification of that mitigation is recorded. - No patch SLA evidenced: neither
SECURITY.mdnor the repository docs state one, so alert ages are recorded but not judged against a standard. - Dependabot security-update status unverified: two security-settings endpoints returned HTTP 403 with this token.
Key metrics:
- Open Dependabot alerts: 2 (1 high, 1 medium). Open CodeQL alerts: 9 (8 high, 1 medium). Open secret-scanning alerts: 0. Open security issues: 0.
- Private vulnerability reporting: enabled. Ownership: one wildcard CODEOWNERS rule naming two individual accounts.
Scope and evidence
Window: 2026-11-05T06:44:41Z to 2026-11-12T06:44:43Z (UTC), checkout d4bf131. The window has 81 commits. The prefetch flags 50 as security-relevant.
Observed:
- The GOV.UK guidance was fetched this run (HTTP 200, published 14 May 2026).
- The repository is public, MIT, not archived, and not a fork.
CODEOWNERSandSECURITY.mdare present. - The private vulnerability reporting endpoint reports enabled.
- A direct paginated Dependabot read (
state=open) returned 2 alerts. The prefetch's code-scanning data lists 9. Secret scanning and security issues returned 0. - Recent changes include dependency manifest updates (root
go.mod4 commits,server/go.mod4,package.json11,package-lock.json3, and an.npmrcfeed change), an Azure Container Apps topology, a Copilot token broker permission change (68679829), and a 31-file removal of committed artifacts, including a SQLite file (62f179a).
Source discrepancy: the prefetch recorded dependabot_alerts as HTTP 400. A direct read in this run returned HTTP 200. Issue #20594 reports that the prefetch sends an unsupported page parameter, which is consistent with this result. This advisory uses the direct read. A previous run reported an incomplete result (issue #20619).
Unverified or inaccessible:
- The Dependabot
vulnerability-alertsandautomated-security-fixessettings returned HTTP 403. The repositorysecurity_and_analysisfields are null, so they establish neither enabled nor disabled. - Branch protection and required status checks were not read.
.github/dependabot.ymlis absent from the checkout and from all fetched commits. Dependabot-style bump commits appear in the window, and their automation source is unconfirmed.- Push protection and history scanning after the artifact removal are not evidenced. Removing files from the tree does not purge history.
Outside the repository: named owner succession, risk-acceptance records, SLA commitments, reachability analysis for alert 103, incident response and on-call, deployed-service telemetry, and rollback drills.
Asset graph
| Surface | Window signal | Ownership signal | Provisional concern |
|---|---|---|---|
Root Go module (go.mod, go.sum) |
Dependabot alerts 103 (high) and 102 (medium); 4 go.mod and 3 go.sum changes |
Wildcard CODEOWNERS only | Fixable high dependency alert; unpatched medium alert |
.github/workflows/release-promote.yml |
CodeQL alerts 39–44 (high, 25 days) | Wildcard CODEOWNERS only | Release-path cache-poisoning findings with an unverified mitigation |
Azure Container Apps topology and server/ |
Topology commits (e77ccf7, 1c66114, 38bc84e); 4 server/go.mod and 4 server/go.sum changes |
Wildcard CODEOWNERS only | Deployment surface changed; no rollback procedure found |
| CAO workflows with the Copilot token broker | 68679829 permission fix; 9d0f85a source-review gate |
Wildcard CODEOWNERS only | Credential-adjacent change; diffs not reviewed in this run |
scripts/validate-agent-docs.mjs and tests |
CodeQL alerts 70 (high, new), 62 (high), and 68 (medium) | Wildcard CODEOWNERS only | Possible false positives; unverified |
Proposed tier classification
Repository-level proposal: B, Open With Conditions candidate. The basis is a fixable high dependency alert, aged high code-scanning alerts with no evidenced SLA, and unverified security-update status. A is excluded by the open high alerts. C and D are not supported, because no exploitation, secret exposure, or compromise evidence was observed. No closure is recommended.
| Surface | Proposed tier | Confidence |
|---|---|---|
| Root Go module | B | Medium |
release-promote.yml |
B | Medium |
Deployment topology and server/ |
B | Low |
| Token broker workflow changes | B | Low |
| Validator and test findings | B | Low |
Control verification gaps
| Control | Status | Evidence | Gap |
|---|---|---|---|
| Clear ownership | Partial | Wildcard CODEOWNERS with two individual accounts | No team or service owner; no succession evidence |
| Secure-by-design development | Partial | Scan-fix commits (d5a1ea9, 896a49f); artifact removal (62f179a) |
No threat-model or hardening evidence in this run |
| Automated dependency and vulnerability hygiene | Partial | Dependabot alerts readable; 2 open | Security-update status unverified (HTTP 403); no dependabot.yml; alert 102 unpatched |
| Patch SLA and remediation capability | Not evidenced | No SLA in SECURITY.md or the docs |
UK-AI-008 is marked implemented, but its own note requires an evidenced SLA |
| Inbound vulnerability reporting | Partial | Private vulnerability reporting enabled; SECURITY.md route |
Generic GitHub mailbox; no project triage or response time |
| Secret exposure | Partial | 0 open secret-scanning alerts | Push protection unknown; history scan not evidenced |
| Runtime observability | Not evidenced | Commit titles only (68679829) |
No alerting, dashboards, or on-call route in the evidence |
| Recovery | Partial | Immutable prerelease tags; release-promote workflow | No rollback runbook or drill evidence |
Risk Scoring and Rationale
Proposed scorer values for the repository: exposure amplification 4, patchability 3, detectability 4, operational fragility 4, and ownership confidence 2 (low). These values are advisory and not calibrated.
The open high alerts exclude A. The fixable transitive dependency alert and the aged cache-poisoning alerts support B. Patchability is reduced by the unpatched alert 102 and by the unknown reachability of alert 103. Ownership confidence is low because the owners are individual accounts and no SLA exists.
Prioritized Remediation Queue
Remediation queue
| # | Action | Urgency | Validation evidence | Owner or gap | Review trigger |
|---|---|---|---|---|---|
| 1 | Update the root go.mod to klauspost/compress v1.18.7 or later (alert 103) |
High | go list -m reports 1.18.7 or later; root build and tests pass; alert closed on re-scan |
Named maintainer not evidenced | Pull request merge, or a blocked upgrade |
| 2 | Verify whether read-only cache mode blocks the path, then fix or dismiss alerts 39–44 with a reviewed rationale | High | Re-scan closes the alerts, or a reviewed dismissal is recorded | No release-workflow owner evidenced | Before the next release promotion |
| 3 | Triage high CodeQL alerts 70 and 62 in validator and test code | Medium | Fix, or a reviewed false-positive dismissal | Owner gap | Next code-scanning review |
| 4 | Record a reviewed risk decision for Dependabot alert 102 (no patch; 2.8.3 and earlier), or remove the dependency. Existing planner issue #20633 tracks it | Medium | Reachability note and acceptance, or removal | No risk-acceptance owner | When a patch is released |
| 5 | Triage medium CodeQL alert 68 in test code | Low | Fix, or a dismissal with rationale | Owner gap | Next code-scanning review |
| 6 | Confirm the Dependabot security-update and branch-protection settings with admin read access | Medium | Settings readback recorded | Admin access not available to this run | Next advisory run |
| 7 | Publish a project-specific patch SLA and triage route, and reconcile UK-AI-008 | Medium | Published SLA and intake text | Maintainer decision | Next advisory run |
| 8 | Document and exercise a rollback runbook for the deployed topology | Medium | Runbook path and drill record | Owner gap | After the next deployment change |
Open-Code Exception Register
None. No closure is recommended. The repository is public, so the private-visibility condition does not apply.
Operational Metrics Baseline
Metrics baseline
| Metric | Observed | Baseline note |
|---|---|---|
| Open alert age vs patch SLA | High alerts: Dependabot 1 (under 1 day); CodeQL 8, of which 7 are 9–25 days old | No SLA evidenced, so a breach cannot be computed |
| Inbound reporting route | Private vulnerability reporting enabled; SECURITY.md names a mailbox |
No project triage process or response time |
| Ownership coverage | One wildcard CODEOWNERS rule (two individual accounts) | Covers all paths; no team or service owner |
| Unsupported dependency ratio | Not measured | No ecosystem-support data was collected in this run |
| Exception aging | 0 exceptions | Not applicable |
| Exposure without recovery capability | Not measured | Immutable prerelease tags exist; no rollback runbook found |
Human Review Required
- CODEOWNERS-listed maintainers: decide whether read-only cache mode mitigates the
release-promote.ymlalerts, or whether a workflow fix is needed. - The same maintainers: decide whether to accept or remove the unpatched
docker/distributiondependency (alert 102) until a patch exists. - Repository admin: confirm the Dependabot security-update and branch-protection settings, which this token cannot read.
- Maintainers: adopt a patch SLA and a project triage route, then reconcile UK-AI-008.
- Reviewer: check this tier and the GOV.UK interpretation against current guidance and organizational context. This advisory does not authorize any change to repository visibility.
Recommended Next Action
The highest expected return is the root go.mod bump of github.com/klauspost/compress to v1.18.7 or later. It clears the only high-severity dependency alert, changes one module pin, and can be verified with the Go build, the tests, and the alert state. The release-promote.yml alerts carry more consequence but need a mitigation decision first, so they follow.
Agent prompt
In githubnext/gh-cao, update the root Go module so it requires github.com/klauspost/compress v1.18.7 or later. This resolves Dependabot alert 103 (GHSA-259r-337f-4rfw / CVE-2026-63209; affected versions 1.18.6 and earlier; first patched version 1.18.7). The affected surface is the root go.mod and go.sum, which currently pin v1.18.0 as an indirect runtime dependency. Required outcome: the root module requires github.com/klauspost/compress at v1.18.7 or later, go.sum is consistent with that requirement, and no other module versions change unless the Go toolchain requires it. Constraints: change only the root go.mod and go.sum; do not edit server/go.mod, any .lock.yml file, any workflow, or any alert state; do not add, remove, or dismiss dependencies or alerts. Verification: go list -m github.com/klauspost/compress reports v1.18.7 or later; go mod verify passes; go build ./... and go test ./... pass from the repository root, or any failure is reported with its cause and not worked around. Include the diff and the verification output in the pull request.
cao: githubnext/gh-cao, correlation: 43169608490-1373
Generated by 🛡️ UK AI Advisory / Resilience · pi · routed: haiku55 xhigh · ~15 AIC · ⌖ ~2 AIC · ⊞ 9.8K · ◷
- expires on Dec 12, 2026, 6:54 AM UTC
- 主要言語
- JavaScript
- スター
- 19
- フォーク
- 4
- 平均マージ
- 4時間
- マージ済み PR(30日)
- 708
環境構築
- Dockerfile・Docker Compose ファイルなし
- プルリクエストのテンプレートなし
- コントリビューションガイドを読む
はじめの一歩
- issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
- 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
- リポジトリをフォークし、ブランチを切って変更します。
- issue 番号を参照したプルリクエストを送ります。
githubnext/gh-cao のほかの issue
-
self-care self-care:agent-discoverability
難易度 2/5 1〜3時間 初心者へのやさしさ 62/100
githubnext/gh-cao#20928 ·
メンテナーはふだん 1 日以内に返信
-
dependabot
難易度 2/5 1〜3時間 初心者へのやさしさ 68/100
githubnext/gh-cao#20776 · コメント 1 件 ·
メンテナーはふだん 1 日以内に返信
-
難易度 2/5 1〜3時間 初心者へのやさしさ 66/100
githubnext/gh-cao#20705 ·
メンテナーはふだん 1 日以内に返信
-
repo-assist repo-assist:maintenance
難易度 1/5 1〜3時間 初心者へのやさしさ 75/100
githubnext/gh-cao#20704 ·
メンテナーはふだん 1 日以内に返信
-
self-care self-care:documentation-discoverability
難易度 2/5 1〜3時間 初心者へのやさしさ 72/100
githubnext/gh-cao#20667 ·
メンテナーはふだん 1 日以内に返信
githubnext/gh-cao の issue をすべて見る
似ている issue
-
framework/gatsby help wanted kind/bug
難易度 2/5 1〜3時間 初心者へのやさしさ 72/100
メンテナーはふだん 1 日以内に返信
-
good first issue help wanted javascript
難易度 2/5 1〜3時間 初心者へのやさしさ 72/100
dlq/synthpopcan#58 ·
メンテナーはふだん 2 日以内に返信
-
cli: `execCargoMetadata` decodes stdout per chunk, corrupting non-ASCII paths across pipe boundariesオープン
難易度 2/5 1〜3時間 初心者へのやさしさ 68/100
メンテナーはふだん 1 日以内に返信
-
難易度 1/5 1時間未満 初心者へのやさしさ 68/100
osmlab/name-suggestion-index#13104 ·
メンテナーはふだん 1 日以内に返信
-
bug triage
難易度 2/5 1〜3時間 初心者へのやさしさ 72/100
fjaeckel/seriesoftubes#376 ·
メンテナーはふだん 1 日以内に返信