Skip to content

Expand live shell approval regression corpus #1930

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Aaronontheweb merged 1 commit into from
Aug 13, 2026
Merged

Expand live shell approval regression corpus #1930

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions IMPLEMENTATION_PLAN.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -205,6 +205,8 @@ Done when:
policy match handles longer parser phrases without private command grammar.
- [x] A synthetic workload corpus covers ordinary search, read, pipeline,
redirect, and file-change commands without production command text.
- [x] Sanitized post-swap evidence classifies 51 prompts across 202 shell calls.
Eleven executable live cases pin the intended allow and prompt boundaries.
- [x] A safe pipeline stage can compose with a stored grant for each stage that
still requires approval.
- [x] A prompt excludes a safe stage from the approval candidates that the user
Expand Down
189 changes: 189 additions & 0 deletions openspec/changes/structure-shell-approval-policy/evidence/netclaw-policy-fixtures.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -410,6 +410,195 @@
"expectedFinal": { "outcome": "RequiresApproval", "reason": "UncoveredCandidates", "approvalCandidates": ["gh pr close"], "isMessy": false, "agentCorrection": null }
}
],
"liveRegressionCases": [
{
"sourceEvidenceId": "S18",
"classification": "NetclawPolicyDebt",
"targetOutcome": "Allow",
"policyCase": {
"id": "L01",
"category": "BoundedGrepAndList",
"command": "grep -Rni 'TypeMarker' src | head; ls src/Observability",
"environment": { "platform": "Linux", "executablePath": "/bin/bash", "commandArguments": ["-c"], "grammar": "Bash", "pathStyle": "Posix", "powerShellDialect": null },
"initialWorkingDirectory": "/work",
"projectDirectory": "/work",
"sessionDirectory": "/work",
"available": { "oneTimeApprovalKeys": [], "sessionGrants": [], "persistentGrants": [], "safePhrases": [] },
"useBundledSafeCatalog": true,
"expected": { "outcome": "Allow", "denyReason": null, "approvalCandidates": null, "isMessy": null, "optionKeys": null, "actorCheckCount": 1 }
}
},
{
"sourceEvidenceId": "S22",
"classification": "NetclawPolicyDebt",
"targetOutcome": "Allow",
"policyCase": {
"id": "L02",
"category": "BoundedGrepPipeline",
"command": "grep -Rni 'TestMarker' tests | head",
"environment": { "platform": "Linux", "executablePath": "/bin/bash", "commandArguments": ["-c"], "grammar": "Bash", "pathStyle": "Posix", "powerShellDialect": null },
"initialWorkingDirectory": "/work",
"projectDirectory": "/work",
"sessionDirectory": "/work",
"available": { "oneTimeApprovalKeys": [], "sessionGrants": [], "persistentGrants": [], "safePhrases": [] },
"useBundledSafeCatalog": true,
"expected": { "outcome": "Allow", "denyReason": null, "approvalCandidates": null, "isMessy": null, "optionKeys": null, "actorCheckCount": 1 }
}
},
{
"sourceEvidenceId": "S40",
"classification": "NetclawPolicyDebt",
"targetOutcome": "Allow",
"policyCase": {
"id": "L03",
"category": "BoundedMultiRootGrep",
"command": "grep -Rni 'PolicyMarker' src tests | head",
"environment": { "platform": "Linux", "executablePath": "/bin/bash", "commandArguments": ["-c"], "grammar": "Bash", "pathStyle": "Posix", "powerShellDialect": null },
"initialWorkingDirectory": "/work",
"projectDirectory": "/work",
"sessionDirectory": "/work",
"available": { "oneTimeApprovalKeys": [], "sessionGrants": [], "persistentGrants": [], "safePhrases": [] },
"useBundledSafeCatalog": true,
"expected": { "outcome": "Allow", "denyReason": null, "approvalCandidates": null, "isMessy": null, "optionKeys": null, "actorCheckCount": 1 }
}
},
{
"sourceEvidenceId": "S16",
"classification": "AgentAlignmentDebt",
"targetOutcome": "RequiresApproval",
"policyCase": {
"id": "L04",
"category": "BroadFindDiscovery",
"command": "find . -maxdepth 3 -name '*.csproj'",
"environment": { "platform": "Linux", "executablePath": "/bin/bash", "commandArguments": ["-c"], "grammar": "Bash", "pathStyle": "Posix", "powerShellDialect": null },
"initialWorkingDirectory": "/work",
"projectDirectory": "/work",
"sessionDirectory": "/work",
"available": { "oneTimeApprovalKeys": [], "sessionGrants": [], "persistentGrants": [], "safePhrases": [] },
"useBundledSafeCatalog": true,
"expected": { "outcome": "RequiresApproval", "denyReason": null, "approvalCandidates": ["find"], "isMessy": false, "optionKeys": ["approve_once", "approve_session", "approve_everywhere", "deny"], "actorCheckCount": 1 }
}
},
{
"sourceEvidenceId": "S11",
"classification": "AgentAlignmentDebt",
"targetOutcome": "RequiresApproval",
"policyCase": {
"id": "L05",
"category": "DiskSummaryWithSort",
"command": "du -sh ./* | sort -rh | head",
"environment": { "platform": "Linux", "executablePath": "/bin/bash", "commandArguments": ["-c"], "grammar": "Bash", "pathStyle": "Posix", "powerShellDialect": null },
"initialWorkingDirectory": "/work",
"projectDirectory": "/work",
"sessionDirectory": "/work",
"available": { "oneTimeApprovalKeys": [], "sessionGrants": [], "persistentGrants": [], "safePhrases": [] },
"useBundledSafeCatalog": true,
"expected": { "outcome": "RequiresApproval", "denyReason": null, "approvalCandidates": ["sort"], "isMessy": false, "optionKeys": ["approve_once", "approve_session", "approve_everywhere", "deny"], "actorCheckCount": 1 }
}
},
{
"sourceEvidenceId": "S10",
"classification": "ExpectedApproval",
"targetOutcome": "RequiresApproval",
"policyCase": {
"id": "L06",
"category": "RemotePullRequestCreate",
"command": "gh pr create --repo example/project --title 'Example' --body 'Example body'",
"environment": { "platform": "Linux", "executablePath": "/bin/bash", "commandArguments": ["-c"], "grammar": "Bash", "pathStyle": "Posix", "powerShellDialect": null },
"initialWorkingDirectory": "/work",
"projectDirectory": "/work",
"sessionDirectory": "/work",
"available": { "oneTimeApprovalKeys": [], "sessionGrants": [], "persistentGrants": [], "safePhrases": [] },
"useBundledSafeCatalog": true,
"expected": { "outcome": "RequiresApproval", "denyReason": null, "approvalCandidates": ["gh pr create"], "isMessy": false, "optionKeys": ["approve_once", "approve_session", "approve_everywhere", "deny"], "actorCheckCount": 1 }
}
},
{
"sourceEvidenceId": "S24",
"classification": "ExpectedApproval",
"targetOutcome": "RequiresApproval",
"policyCase": {
"id": "L07",
"category": "RemotePullRequestMerge",
"command": "gh pr merge 50 --repo example/project --auto --squash",
"environment": { "platform": "Linux", "executablePath": "/bin/bash", "commandArguments": ["-c"], "grammar": "Bash", "pathStyle": "Posix", "powerShellDialect": null },
"initialWorkingDirectory": "/work",
"projectDirectory": "/work",
"sessionDirectory": "/work",
"available": { "oneTimeApprovalKeys": [], "sessionGrants": [], "persistentGrants": [], "safePhrases": [] },
"useBundledSafeCatalog": true,
"expected": { "outcome": "RequiresApproval", "denyReason": null, "approvalCandidates": ["gh pr merge"], "isMessy": false, "optionKeys": ["approve_once", "approve_session", "approve_everywhere", "deny"], "actorCheckCount": 1 }
}
},
{
"sourceEvidenceId": "S20",
"classification": "ExpectedApproval",
"targetOutcome": "RequiresApproval",
"policyCase": {
"id": "L08",
"category": "RemoteGitPush",
"command": "git push origin feature/example",
"environment": { "platform": "Linux", "executablePath": "/bin/bash", "commandArguments": ["-c"], "grammar": "Bash", "pathStyle": "Posix", "powerShellDialect": null },
"initialWorkingDirectory": "/work",
"projectDirectory": "/work",
"sessionDirectory": "/work",
"available": { "oneTimeApprovalKeys": [], "sessionGrants": [], "persistentGrants": [], "safePhrases": [] },
"useBundledSafeCatalog": true,
"expected": { "outcome": "RequiresApproval", "denyReason": null, "approvalCandidates": ["git push origin"], "isMessy": false, "optionKeys": ["approve_once", "approve_session", "approve_everywhere", "deny"], "actorCheckCount": 1 }
}
},
{
"sourceEvidenceId": "S13",
"classification": "ExpectedApproval",
"targetOutcome": "RequiresApproval",
"policyCase": {
"id": "L09",
"category": "RecursiveDelete",
"command": "rm -rf ./cache/*",
"environment": { "platform": "Linux", "executablePath": "/bin/bash", "commandArguments": ["-c"], "grammar": "Bash", "pathStyle": "Posix", "powerShellDialect": null },
"initialWorkingDirectory": "/work",
"projectDirectory": "/work",
"sessionDirectory": "/work",
"available": { "oneTimeApprovalKeys": [], "sessionGrants": [], "persistentGrants": [], "safePhrases": [] },
"useBundledSafeCatalog": true,
"expected": { "outcome": "RequiresApproval", "denyReason": null, "approvalCandidates": ["rm"], "isMessy": false, "optionKeys": ["approve_once", "approve_session", "approve_everywhere", "deny"], "actorCheckCount": 1 }
}
},
{
"sourceEvidenceId": "S21",
"classification": "ExpectedApproval",
"targetOutcome": "RequiresApproval",
"policyCase": {
"id": "L10",
"category": "DynamicDiagnosticRange",
"command": "dotnet test > /tmp/test.log; sed -n \"$(grep -n 'FAIL' /tmp/test.log | cut -d: -f1),+4p\" /tmp/test.log",
"environment": { "platform": "Linux", "executablePath": "/bin/bash", "commandArguments": ["-c"], "grammar": "Bash", "pathStyle": "Posix", "powerShellDialect": null },
"initialWorkingDirectory": "/work",
"projectDirectory": "/work",
"sessionDirectory": "/work",
"available": { "oneTimeApprovalKeys": [], "sessionGrants": [], "persistentGrants": [], "safePhrases": [] },
"useBundledSafeCatalog": true,
"expected": { "outcome": "RequiresApproval", "denyReason": null, "approvalCandidates": [], "isMessy": true, "optionKeys": ["approve_once", "deny"], "actorCheckCount": 0 }
}
},
{
"sourceEvidenceId": "S44",
"classification": "NetclawPolicyDebt",
"targetOutcome": "Allow",
"policyCase": {
"id": "L11",
"category": "BoundedResourceGrep",
"command": "grep -Rni 'ResourceMarker' src | head",
"environment": { "platform": "Linux", "executablePath": "/bin/bash", "commandArguments": ["-c"], "grammar": "Bash", "pathStyle": "Posix", "powerShellDialect": null },
"initialWorkingDirectory": "/work",
"projectDirectory": "/work",
"sessionDirectory": "/work",
"available": { "oneTimeApprovalKeys": [], "sessionGrants": [], "persistentGrants": [], "safePhrases": [] },
"useBundledSafeCatalog": true,
"expected": { "outcome": "Allow", "denyReason": null, "approvalCandidates": null, "isMessy": null, "optionKeys": null, "actorCheckCount": 1 }
}
}
],
"adversarialCases": [
{
"id": "A01",
Expand Down
Loading
Loading