feat(policy): MVP PolicyEvaluator (closes #21)#55
Merged
Conversation
- Add PolicyEvaluator: CheckRun (wall clock, cost), CheckStep (structured output), CheckToolCall (forbidUnknownTools, approval-required actions) - Split helpers: budget, permissions (uses tools.ParseUses), approvals - DeniedError with TraceData for trace.EventPolicyDenied - Engine resolves named Policy resources from the project graph RunContext carries elapsed time, accumulated cost, and approved actions (for future repeated --approve on agentctl run). Closes #21 Made-with: Cursor
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Implements the policy engine MVP from issue #21 and design doc section 12.2 H: run/step/tool checks for wall-clock budget, cost ceiling, structured output requirement, unknown tools, and approval-gated tool actions.
API
CheckRun,CheckStep,CheckToolCall(matches design doc interface).Elapsed,AccumulatedCostUSD,ApprovedActions(slice of fullusesstrings, intended to be filled from repeated--approve <action>whenagentctl runexists).Evaluator(policyName)resolvesPolicyspecs from the graph;EvaluatorForSpecfor explicit merged specs (e.g. env overrides later).reasonconstants +TraceData()fortrace.EventPolicyDenied.Helpers (as requested)
budget.gopermissions.goforbidUnknownToolsviatools.ParseUses+ graphToolsapprovals.goapprovals.requiredForexact match onusesvsApprovedActionsTests
forbidUnknownTools: true→*DeniedError(unknown_tool).useswithout a matching approved action → denial;TraceDataappended aspolicy.deniedvia SQLite +trace.Recorder, then read back.--approve-style slice allows call, wall clock / cost / structured output denials,Engine.Evaluatornamed policy resolution.Verification
make fmtmake vetmake test(-race)Notes
runis not wired yet;RunContext.ApprovedActionsdefines the contract for repeated--approvevalues matching YAMLrequiredForentries (e.g.tool.github.pull_request.merge).Closes #21
Made with Cursor