ctxora 6.3.0 → 6.4.0

package/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "ctxora",
-  "version": "6.3.0",
+  "version": "6.4.0",
   "description": "CTXORA repository context, memory, handoff, and diagnostics workflows.",
   "homepage": "https://ctxora-landing.vercel.app/",
   "repository": "https://github.com/nguyentrunghieutcu/ctxora-engine",

package/README.md

@@ -180,10 +180,12 @@ npx ctxora skills preview --workspace . --profile developer \
   --add-module security --remove-skill security-scan
 npx ctxora skills install --workspace . --profile developer \
   --add-module security --remove-skill security-scan
+ctxora skills install --workspace . --profile developer \
+  --target codex --target claude --target cursor --target gemini
 ctxora skills route "Fix React hydration performance" --workspace . --profile developer
 ```
 
-The default installation directory is `.agents/skills`; use `--output` for another host-specific directory. Preview never mutates files. Install refuses to overwrite modified skills unless `--force` is explicit, and `--prune` removes only unchanged skills previously installed by CTXORA. See `THIRD_PARTY_NOTICES.md` for provenance and licensing.
+The default target is Codex-compatible `.agents/skills`. Use repeatable `--target codex|claude|cursor|gemini|opencode`, `--target all`, or `--output` for a custom directory. Preview never mutates files. Install refuses to overwrite modified skills unless `--force` is explicit, and `--prune` removes only unchanged skills previously installed by CTXORA. See `THIRD_PARTY_NOTICES.md` for provenance and licensing.
 
 ### Commands and agent roles
 

package/README.vi.md

@@ -180,9 +180,11 @@ npx ctxora skills preview --workspace . --profile developer \
   --add-module security --remove-skill security-scan
 npx ctxora skills install --workspace . --profile developer \
   --add-module security --remove-skill security-scan
+ctxora skills install --workspace . --profile developer \
+  --target codex --target claude --target cursor --target gemini
 ```
 
-Thư mục cài mặc định là `.agents/skills`; dùng `--output` cho thư mục skill riêng của host khác. Preview không sửa file. Install không ghi đè skill đã chỉnh sửa nếu chưa truyền `--force`; `--prune` chỉ xóa skill chưa bị sửa và trước đó do CTXORA cài. Xem provenance và license tại `THIRD_PARTY_NOTICES.md`.
+Target mặc định là `.agents/skills` tương thích Codex. Có thể lặp `--target codex|claude|cursor|gemini|opencode`, dùng `--target all`, hoặc `--output` cho thư mục tùy chỉnh. Preview không sửa file. Install không ghi đè skill đã chỉnh sửa nếu chưa truyền `--force`; `--prune` chỉ xóa skill chưa bị sửa và trước đó do CTXORA cài. Xem provenance và license tại `THIRD_PARTY_NOTICES.md`.
 
 ### Command và agent role
 
@@ -193,6 +195,7 @@ CTXORA cũng cung cấp các template workflow chủ động, lấy cảm hứng
 | Command | Mục đích | Khả năng CTXORA |
 |---|---|---|
 | `/ctxora:context <task>` | Lấy bằng chứng repository trước khi code | `plan_context`, `retrieve_context`, `prepare_context` |
+| `/ctxora:route <task>` | Tự chọn skill trong profile và học sau validation | `route_skills`, `skill_feedback` |
 | `/ctxora:plan <task>` | Lập kế hoạch dựa trên context | `plan_context`, `retrieve_context` |
 | `/ctxora:review [scope]` | Review thay đổi với context repository | `retrieve_context` |
 | `/ctxora:health` | Kiểm tra workspace và index | `doctor`, `context-score`, `inspect` |

package/bin/ctxora.mjs

@@ -6,7 +6,7 @@ import { dirname, join, resolve } from "node:path";
 import { fileURLToPath } from "node:url";
 import { spawnSync } from "node:child_process";
 
-const PACKAGE_VERSION = "6.3.0";
+const PACKAGE_VERSION = "6.4.0";
 const PYTHON_RANGE = "3.10-3.13";
 const PACKAGE_ROOT = resolve(dirname(fileURLToPath(import.meta.url)), "..");
 

package/commands/route.md

@@ -3,4 +3,4 @@ description: Automatically rank the installed ECC skills for a task.
 argument-hint: <task>
 ---
 
-Call `route_skills` for the registered workspace and `$ARGUMENTS`. Use the highest-ranked relevant recommendations, read their `SKILL.md` files before acting, and retain the returned `route_id`. After validation, call `skill_feedback` with the skills actually used and the outcome. Do not execute bundled scripts or external actions merely because a routed skill mentions them.
+Call `route_skills` for the registered workspace and `$ARGUMENTS`. Use the highest-ranked relevant recommendations, read their `SKILL.md` files before acting, and retain the returned `route_id` in task state. After validation passes, automatically call `skill_feedback` with outcome `success` and the skills actually used. Use `failure` only when the failed result is attributable to the guidance; if the user replaces the selected skill, use `corrected`. Do not ask the user to submit routine feedback manually, and do not execute bundled scripts or external actions merely because a routed skill mentions them.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "ctxora",
-  "version": "6.3.0",
+  "version": "6.4.0",
   "description": "Local-first context engine for coding agents.",
   "type": "module",
   "bin": {

package/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "ctxora-engine"
-version = "6.3.0"
+version = "6.4.0"
 description = "Local-first context engine for coding agents."
 readme = "README.md"
 requires-python = ">=3.10"

package/skills/ctxora-continuous-learning/SKILL.md

@@ -7,7 +7,7 @@ description: Capture evidence-backed project lessons in CTXORA memory and promot
 
 CTXORA learning is deliberate and memory-based; it does not require background hooks.
 
-Skill-routing learning is separate from durable memory. Keep the `route_id` returned by `route_skills` or `prepare_context`, then call `skill_feedback` after validation. CTXORA stores only a task fingerprint, matched catalog terms, bounded counters, and decayed associations; it never stores raw tasks or transcripts.
+Skill-routing learning is separate from durable memory. Keep the `route_id` returned by `route_skills` or `prepare_context` in task state. After validation passes, automatically call `skill_feedback` with `success` and the skills actually used; use `failure` only when the failed result is attributable to the guidance, and `corrected` when the user replaces the routed skill. Do not ask the user to submit routine feedback manually. CTXORA stores only a task fingerprint, matched catalog terms, bounded counters, and decayed associations; it never stores raw tasks or transcripts.
 
 Inspect connected tool schemas and resolve the registered workspace ID before using `memory_search` or `memory_save`. If tools are unavailable, present a proposed lesson and explicitly report that it was not saved. Do not write directly into memory databases or ECC vaults. Do not capture raw session transcripts or install observers as part of this skill.
 

package/src/harness_context/application/context_service.py

@@ -47,7 +47,7 @@ class ContextService:
         if self.skill_router is not None:
             try:
                 routed = self.skill_router.route(request.workspace_id, request.query, top_k=3, include_instructions=False, token_budget=0)
-                skill_diagnostics.update({"route_id": routed["route_id"], "profile": routed["profile"], "recommendations": routed["recommendations"]})
+                skill_diagnostics.update({"route_id": routed["route_id"], "profile": routed["profile"], "recommendations": routed["recommendations"], "feedback_policy": routed["feedback_policy"]})
             except ValueError as error:
                 skill_diagnostics.update({"recommendations": [], "routing_error": str(error)})
         elapsed_ms = (time.perf_counter() - started) * 1000

package/src/harness_context/cli/app.py

@@ -119,7 +119,13 @@ def build_parser() -> argparse.ArgumentParser:
         action_parser.add_argument("--remove-module", action="append", default=[])
         action_parser.add_argument("--add-skill", action="append", default=[])
         action_parser.add_argument("--remove-skill", action="append", default=[])
-        action_parser.add_argument("--output", default=".agents/skills")
+        action_parser.add_argument("--output", default="")
+        action_parser.add_argument(
+            "--target",
+            action="append",
+            choices=("codex", "claude", "cursor", "gemini", "opencode", "all"),
+            default=[],
+        )
         action_parser.add_argument("--force", action="store_true")
         action_parser.add_argument("--prune", action="store_true")
         if action == "install":
@@ -191,13 +197,25 @@ def _main(argv: list[str] | None = None) -> int:
             tuple(args.remove_skill),
         )
         dry_run = args.skills_action == "preview" or args.dry_run
-        result = catalog.install(root, args.output, selection, force=args.force, prune=args.prune, dry_run=dry_run)
-        result["status"] = "conflict" if result["conflicts"] else ("planned" if dry_run else "installed")
-        result["selection"] = selection.to_dict()
-        if not dry_run and not result["conflicts"]:
+        installs = []
+        conflicts = []
+        for target, output in catalog.install_outputs(tuple(args.target), args.output):
+            installed = catalog.install(
+                root, output, selection, force=args.force, prune=args.prune, dry_run=dry_run
+            )
+            installed["target"] = target
+            installs.append(installed)
+            conflicts.extend(f"{target}:{skill}" for skill in installed["conflicts"])
+        result = {
+            "status": "conflict" if conflicts else ("planned" if dry_run else "installed"),
+            "selection": selection.to_dict(),
+            "installs": installs,
+            "conflicts": conflicts,
+        }
+        if not dry_run and not conflicts:
             result["profile_config"] = str(catalog.write_selection(root, selection))
         _print(result)
-        return 2 if result["conflicts"] else 0
+        return 2 if conflicts else 0
     if args.command == "setup":
         config = workspace_data_dir(root) / "config.toml"
         config.parent.mkdir(parents=True, exist_ok=True)

package/src/harness_context/skills/catalog.py

@@ -9,6 +9,14 @@ from dataclasses import dataclass
 from pathlib import Path
 from typing import Any
 
+SKILL_INSTALL_TARGETS = {
+    "codex": ".agents/skills",
+    "claude": ".claude/skills",
+    "cursor": ".cursor/skills",
+    "gemini": ".gemini/skills",
+    "opencode": ".opencode/skills",
+}
+
 
 @dataclass(frozen=True)
 class SkillSelection:
@@ -95,6 +103,20 @@ class SkillCatalog:
     def source_root(self) -> Path:
         return Path(__file__).with_name("ecc")
 
+    @staticmethod
+    def install_outputs(targets: tuple[str, ...], output: str = "") -> tuple[tuple[str, str], ...]:
+        if output and targets:
+            raise ValueError("use either --output or --target, not both")
+        if output:
+            return (("custom", output),)
+        selected = list(targets or ("codex",))
+        if "all" in selected:
+            selected = list(SKILL_INSTALL_TARGETS)
+        unknown = sorted(set(selected) - set(SKILL_INSTALL_TARGETS))
+        if unknown:
+            raise ValueError(f"unknown skill install target: {unknown[0]}")
+        return tuple((name, SKILL_INSTALL_TARGETS[name]) for name in dict.fromkeys(selected))
+
     def install(
         self,
         workspace: str | Path,

package/src/harness_context/skills/router.py

@@ -165,6 +165,12 @@ class SkillRouter:
             "candidate_count": len(candidates),
             "recommendations": recommendations,
             "learning": {"project_scoped": True, "raw_tasks_stored": False},
+            "feedback_policy": {
+                "mode": "after_validation",
+                "retain_route_id": True,
+                "success_requires_validation": True,
+                "user_correction_outcome": "corrected",
+            },
         }
 
     def feedback(