@hank-warren/pi-auto-permissions 0.6.0 → 0.7.0

package/CHANGELOG.md

@@ -0,0 +1,38 @@
+# @hank-warren/pi-auto-permissions
+
+## 0.7.0
+
+### Minor Changes
+
+- ad12ad0: Rename the eval-logging prompt's block option to "Block — asking was appropriate".
+
+  Label clarity only — the schema is unchanged (`userChoice: "block"`, `expectedDecision: "ask_user"`). A block always affirms that asking was right: the guardian has no reject verdict (its non-approve outcomes are `ask_user` and `revise`), so the only true rejection in the system is the user's at this prompt. Plain "Block" from the logging-disabled prompt still classifies identically, including for override evidence.
+
+## 0.6.1
+
+### Patch Changes
+
+- cf12677: ship CHANGELOG.md in the published tarball
+- Updated dependencies [cf12677]
+  - @hank-warren/pi-permission-selector@1.0.1
+
+## 0.6.0
+
+### Minor Changes
+
+- 500fe67: Remove the `ExtensionSelectorComponent` monkey patch; compose `OptionSelector` explicitly (spec §9.4).
+
+  `pi-permission-selector` is now a **pure library**: it no longer registers an extension, patches nothing, and standalone installs do nothing — the reason for the major bump. `selector-patch.ts` and `index.ts` are deleted; `keys.ts` and `selector.ts` remain the frozen cross-package surface. Dialogs this repo does not own (pi's model/theme/session pickers, third-party extensions) revert to stock arrows+Enter, an accepted trade-off (spec §9.3): nothing here breaks when a pi upgrade changes selector internals.
+
+  `pi-auto-permissions` renders both of its prompts — the guardian approval prompt and the `request_override` prompt — through `ctx.ui.custom()` with the shared `OptionSelector` from `@hank-warren/pi-permission-selector` (a new plain dependency), instead of `ctx.ui.select` plus the global patch. Behavior is preserved: numbered options with `1`–`9` hotkeys, Tab-to-comment on the approval prompt (the note is still delivered as a steering user message, now by pi-auto-permissions itself), exact legacy option strings for `classifyPromptChoice`, and Herdr blocked-state bracketing. Cancellation semantics match `ctx.ui.select`: Esc, abort of the prompt signal, and a host that cannot render custom UI all resolve `undefined`, which every caller treats as deny/cancel — never allow.
+
+### Patch Changes
+
+- Updated dependencies [500fe67]
+  - @hank-warren/pi-permission-selector@1.0.0
+
+## 0.5.1
+
+### Patch Changes
+
+- 67abcf5: document the optional `HERDR_ENV` herdr pane blocked indicator

package/README.md

@@ -201,7 +201,7 @@ The path defaults to `review-evals.jsonl` beside the Auto Permissions config and
 When logging is enabled, prompted reviews offer three choices:
 
 - **Allow — asking was unnecessary** executes the command and records `userChoice: "allow_unnecessary"` with `expectedDecision: "approve"`.
-- **Block** remains the second choice, blocks the command, and records `userChoice: "block"` with `expectedDecision: "ask_user"`.
+- **Block — asking was appropriate** remains the second choice, blocks the command, and records `userChoice: "block"` with `expectedDecision: "ask_user"`. (A block always affirms the prompt: the guardian has no reject verdict — its only non-approve outcomes are asking you or bouncing the command back to the agent as `revise` — so the only true rejection in the system is yours at this prompt.)
 - **Allow — asking was appropriate** executes the command and records `userChoice: "allow_appropriate"` with `expectedDecision: "ask_user"`.
 
 When logging is disabled, the prompt retains the normal **Allow** and **Block** choices. Prompts from Pi or other extensions are unchanged.

package/evaluation-log.ts

@@ -5,7 +5,7 @@ import type { PermissionDecision, ReviewEvidenceRecord } from "./review.js";
 export const PROMPT_FEEDBACK_OPTIONS = {
   allowUnnecessary: "Allow — asking was unnecessary",
   allowAppropriate: "Allow — asking was appropriate",
-  block: "Block",
+  block: "Block — asking was appropriate",
 } as const;
 
 export type PromptEvaluationUserChoice = "allow_unnecessary" | "allow_appropriate" | "block";
@@ -47,6 +47,10 @@ export interface PromptEvaluationRecord {
 
 export function classifyPromptChoice(choice: string | undefined): PromptChoiceClassification | undefined {
   if (choice === "Allow") return { allowsExecution: true };
+  // Plain "Block" is still live: the logging-disabled prompt offers
+  // ["Allow", "Block"]. It carries the same block semantics as the labeled
+  // variant so override evidence works in both prompt modes.
+  if (choice === "Block") return { allowsExecution: false, userChoice: "block" };
   if (choice === PROMPT_FEEDBACK_OPTIONS.allowUnnecessary) {
     return { allowsExecution: true, userChoice: "allow_unnecessary" };
   }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@hank-warren/pi-auto-permissions",
-  "version": "0.6.0",
+  "version": "0.7.0",
   "description": "Fork of @ogulcancelik/pi-auto-permissions with dialog-answer user evidence: context-aware Bash permissions for Pi with automated guardian review.",
   "type": "module",
   "keywords": [
@@ -49,10 +49,11 @@
     "usage-log.ts",
     "widget-status.ts",
     "README.md",
-    "LICENSE"
+    "LICENSE",
+    "CHANGELOG.md"
   ],
   "dependencies": {
-    "@hank-warren/pi-permission-selector": "^1.0.0"
+    "@hank-warren/pi-permission-selector": "^1.0.1"
   },
   "peerDependencies": {
     "@earendil-works/pi-ai": "*",