anthropics
[FEATURE] Allow for updated input in `PermissionRequest:ExitPlanTool:Approve` hook
Preflight Checklist
- [x] I have searched existing requests and this feature hasn't been requested yet
- [x] This is a single feature request (not multiple features)
Problem Statement
I would like to be able to send additional feedback along with approval of a plan with PermissionRequest hook responds with behavior:allow (for tool ExitPlanMode). For instance, accepting a plan but guiding Claude with additional instructions for the plan. This is useful for users who work with spec-driven development.
See a user request here: https://github.com/backnotprop/plannotator/issues/7
Related to #15755
Proposed Solution
Within the hook tool call sequence for approving a plan, iePermissionRequest:ExitPlanTool:Approve, please add the ability to submit extra context.
"hookSpecificOutput": { "hookEventName": "PermissionRequest", "decision": { "behavior": "allow", "updatedInput": { "plan": "{full_plan}", // or: "appendToPlan": "{extra details}" } } } }
Alternative Solutions
No response
Priority
Medium - Would be very helpful
Feature Category
Other
Use Case Example
Plugin, check it out.
https://github.com/backnotprop/plannotator/issues/7
Additional Context
No response
https://github.com/bradennapier/ai-model-registry/tree/main/docs/specs
- Incremental spec driven dev <3
- stateful and resumable
It prob makes most sense to just allow sending a message property so that it works the same way as deny works today.
This may have been completed in v2.1.7 as well. wow big update to close 2 issues.
yes, +1. would be very useful to give feedback on plan- on localhost review
