Repository navigation
docs(adr): add ADR 0161, repo config limits what a run artifact publishes - #9182
Conversation
PR Summary by QodoDocument repo-level limits on published run artifacts
AI Description
Diagram
High-Level Assessment
Files changed (4)
|
Code Review by Qodo
1.
|
|
🤖 Finished Review · ✅ Success · Started 1:34 AM UTC · Completed 1:48 AM UTC Commit: Runtime: pi · Model: sonnet → claude-sonnet-5-5 · Effort: high · Cost: $2.41 |
Site previewPreview: https://214b01a2-site.fullsend-ai.workers.dev Commit: |
|
Risk Assessment: moderate (2/5) DetailsDocumentation-only addition of ADR 0161 with no protected paths or code modifications, elevated to moderate due to high recent commit churn and multi-author contention on touched architecture documentation. Previous runRisk Assessment: moderate (2/5) DetailsDocumentation-only addition of ADR 0161 with no protected paths or code modifications, elevated to moderate due to high recent commit churn and multi-author contention on touched architecture documentation. Previous run (2)Risk Assessment: moderate (2/5) DetailsDocumentation-only addition of ADR 0161 with no protected paths, elevated to moderate due to high recent commit churn and multi-author contention on touched architecture documentation. Previous run (3)Risk Assessment: moderate (2/5) DetailsDocumentation-only addition of ADR 0161 with no protected paths, elevated to moderate due to high recent commit churn and multi-author contention on touched architecture documentation. Previous run (4)Risk Assessment: moderate (2/5) DetailsDocumentation-only addition of ADR 0161 with no protected paths or code modifications, elevated to moderate due to high recent commit churn and multi-author contention on touched architecture documentation. Previous run (5)Risk Assessment: moderate (2/5) DetailsDocumentation-only addition of ADR 0161 with no protected paths or code modifications, elevated to moderate due to high recent commit churn and multi-author contention on touched architecture documentation. |
|
Looks good to me Previous run<!-- fullsend:review-findings-v2:eyJ2ZXJzaW9uIjoyLCJhY3Rpb24iOiJhcHByb3ZlIiwiZmluZGluZ3MiOlt7InNldmVyaXR5IjoibWVkaXVtIiwiY2F0ZWdvcnkiOiJsb2dpYy1lcnJvciIsImZpbGUiOiJkb2NzL0FEUnMvMDE2MS1yZXBvLWNvbmZpZy1saW1pdHMtcnVuLWFydGlmYWN0LWNvbnRlbnQubWQiLCJpZCI6ImZfODI5YzU5ZTI2NDcyODAyNiIsImxpbmUiOjkzLCJhY3Rpb25hYmxlIjp0cnVlLCJkaW1lbnNpb24iOiJjb3JyZWN0bmVzcyJ9LHsic2V2ZXJpdHkiOiJsb3ciLCJjYXRlZ29yeSI6ImluY29tcGxldGUtZG9jIiwiZmlsZSI6ImRvY3MvYXJjaGl0ZWN0dXJlLm1kIiwiaWQiOiJmX2FmZThmMTczMDgwODAyNzUiLCJsaW5lIjo3MTgsImFjdGlvbmFibGUiOnRydWUsImRpbWVuc2lvbiI6ImRvY3MtY3VycmVuY3kifV0sImRpc3Bvc2l0aW9ucyI6W3siaWQiOiJmXzgyOWM1OWUyNjQ3MjgwMjYiLCJzdGF0dXMiOiJyZXNvbHZlZF9ieV9jaGFuZ2UifSx7ImlkIjoiZl9hZmU4ZjE3MzA4MDgwMjc1Iiwic3RhdHVzIjoicmVzb2x2ZWRfYnlfY2hhbmdlIn1dfQ== --> Looks good to me Labels: PR adds an ADR and amends related documentation. Previous run (2)<!-- fullsend:review-findings-v2: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 --> Looks good to me Previous run (3)<!-- fullsend:review-findings-v2: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 --> Looks good to me Earlier findings
Previous run (4)<!-- fullsend:review-findings-v2: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 --> ReviewFindingsMedium
Low
Next steps:
Previous run (5)<!-- fullsend:review-findings-v2:eyJ2ZXJzaW9uIjoyLCJhY3Rpb24iOiJyZXF1ZXN0LWNoYW5nZXMiLCJmaW5kaW5ncyI6W3sic2V2ZXJpdHkiOiJtZWRpdW0iLCJjYXRlZ29yeSI6ImxvZ2ljLWVycm9yIiwiZmlsZSI6ImRvY3MvQURScy8wMTYxLXJlcG8tY29uZmlnLWxpbWl0cy1ydW4tYXJ0aWZhY3QtY29udGVudC5tZCIsImlkIjoiZl84MjljNTllMjY0NzI4MDI2IiwibGluZSI6NzMsImFjdGlvbmFibGUiOnRydWUsImRpbWVuc2lvbiI6ImNvcnJlY3RuZXNzIn1dfQ== --> ReviewFindingsMedium
Next steps:
|
|
🤖 Finished Review · ✅ Success · Started 1:59 AM UTC · Completed 2:11 AM UTC Commit: Runtime: pi · Model: sonnet → claude-sonnet-5-5 · Effort: high · Cost: $1.65 |
There was a problem hiding this comment.
It does makes sense to me, I'm not sure if we want to adopt a more flexible way to apply this: maybe at harness level, overridden by overlays, overridden by config's agents[] entries, overridden by root config.yaml configuration. And the config.yaml would take precedence. Myabe using CEL expressions?
|
🤖 Finished Review · ✅ Success · Started 2:02 PM UTC · Completed 2:09 PM UTC Commit: Runtime: pi · Model: sonnet → claude-sonnet-5-5 · Effort: high · Cost: $1.03 |
Superseded by updated review
|
🤖 Finished Review · Commit: Runtime: pi · Model: sonnet → claude-sonnet-5-5 · Effort: high · Cost: $1.14 This review was superseded by a newer commit. Check the current pull request HEAD's review status before merging. |
Give the telemetry scrub its own bullet, and say that artifact consumers also lose security findings under metadata. Refs #9122 Assisted-by: Claude Signed-off-by: Wayne Sun <gsun@redhat.com>
|
🤖 Finished Review · ✅ Success · Started 8:06 PM UTC · Completed 8:13 PM UTC Commit: Runtime: pi · Model: sonnet → claude-sonnet-5-5 · Effort: high · Cost: $1.25 |
Superseded by updated review
maruiz93
left a comment
There was a problem hiding this comment.
I have two clarification requests about the security boundary and transcript-dependent automation.
…ve no private store Assisted-by: Claude Signed-off-by: Wayne Sun <gsun@redhat.com>
|
🤖 Finished Review · ✅ Success · Started 4:28 PM UTC · Completed 4:34 PM UTC Commit: Runtime: pi · Model: sonnet → claude-sonnet-5-5 · Effort: high · Cost: $1.20 |
Superseded by updated review
Adds ADR 0161 for #9122. It amends ADR 0021: a repo-level
artifactssetting inconfig.yamllimits what a run artifact publishes for each run source, so a public repo can take work from issue trackers that are not public without publishing what its agents read. Jira is the first such source; the setting is not Jira-specific.Decision
alloutput/, as today (default)metadatametrics.json,eval-measurements.jsonl,run-telemetry.jsonlwithout conversation content.security/findings.jsonlstays on the runner, because a finding'sdetailcan quote tool outputnonecontentandby_source[<run source>]wins. Keys are the run source names (github,gitlab,jira,manual,schedule, the same asevent.source.system) and stay open strings for future tracker adapters.jira-roselects the same runs asjira, so one name is enough (recorded under Options).metadatais an allowlist. Filtering builds a staged copy aftereval-measureand post-scripts, and fails closed.allgoes through the ADR 0122 safety gate.Implementation lands separately via #9122 after this merges.
Files
docs/ADRs/0161-repo-config-limits-run-artifact-content.md: new ADRdocs/ADRs/0021-jsonl-reasoning-trace-exposure.md: "Amended by" notedocs/architecture.md,docs/problems/operational-observability.md: cross-referencesEvidence (main at 5a9629d)
action.yml, step "Upload fullsend artifacts": uploads${{ github.workspace }}/outputunderif: always(), with no skip or filter.internal/scaffold/fullsend-repo-gitlab/.gitlab/ci/fullsend-agent.yml:artifacts: paths: [output/],expire_in: 1 week.internal/cli/run.go:iteration-<N>/holdsoutput.jsonl(raw stream),transcripts/andoutput/. Run-level files aremetrics.jsonandsecurity/findings.jsonl.internal/cli/content_collector.go: conversation content entersrun-telemetry.jsonlonly when Level 3 capture is on, and that file is exempt from the host output scan.internal/agentnew/roles.go: retro is the only role with thegithub-artifactsprovider, so it reads past run artifacts.🤖 Generated with Claude Code