pwn 0.5.669 → 0.5.670

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 746a770eb9c9d4d410cd3fff4deb631ac67476a8d4f01ba7cf17c8414d4930d9
-  data.tar.gz: 0aa9628ff2b8f80ed5b06fe682d8c252d4609c3545c96b68dd3890897de3d402
+  metadata.gz: df46a72d0f582458534c1bcfcbdc0fcbf121cf002828ff5da313a5cf5eddea37
+  data.tar.gz: c4bf47cbbe4e7ad22d52104f4fa9af499c06176cae9b93500652956428abe4a2
 SHA512:
-  metadata.gz: adee5114bbd4a83820b12f591d334aeec378f5bf47d7b51d10e3ee9d14497e08207ee532eb2eab7881cb8bb2a1f8910ef6a83712b8b3cb1f7ed7d25cc14f0312
-  data.tar.gz: 5343e693c473baf4112b47b1c8a932a22b6ec46ddf1998c34e828aa79483d7c350bce500ed502f015944960b8e8ec867028f4261f11ddd57360853286778a829
+  metadata.gz: c51242fc7542f675245804852cb313d75a8779675032148585cfde8afab392934d9da03eef08765d1339bca64732c7c3d7350a6ff76a366801c36e5b9e315df5
+  data.tar.gz: 983ff27f8a551664fbc42a5d72357389fee1aaeeb67da5c7b82e4704fa54c06d469aa4589585369c48eca7441193644fd837a08049ee3a560a6bb99f424ba007

data/Gemfile

@@ -18,7 +18,7 @@ gem 'aws-sdk', '3.3.0'
 gem 'barby', '0.7.0'
 gem 'base32', '0.3.4'
 gem 'bitcoin-ruby', '0.0.20'
-gem 'brakeman', '8.0.5'
+gem 'brakeman', '8.0.6'
 gem 'bson', '5.2.0'
 gem 'bundler', '>=4.0.18'
 gem 'bundler-audit', '>=0.9.3'

data/README.md

@@ -118,8 +118,10 @@ message bus:
 ![Swarm Multi-Agent](documentation/diagrams/swarm-multi-agent.svg)
 
 Long-running turns also show **executive task briefs** (not raw commands) via
-`TaskSummarizer`: one full plan on submit (`emit_plan!`), then per-batch
-`about_to` lines keyed by `tool_counts_phrase` + `intent_phrase` with
+`TaskSummarizer`. Every request is first classified as a **general statement**,
+a **question**, or an **autonomous goal**. Only autonomous goals get a multi-step
+breakdown on submit (`emit_plan!`); statements and questions stay single-turn.
+Per-batch `about_to` lines use `tool_counts_phrase` + `intent_phrase` with
 `last_brief_fp` duplicate suppression. When recent turns keep hitting the iteration ceiling, the Loop tightens the remaining runway (lower `max_iters` on local engines, text-only tail, no counterfactual fork) so the agent still finishes instead of thrashing.
 
 Full pages: [How PWN Works](documentation/How-PWN-Works.md) ·

data/documentation/AI-Integration.md

@@ -48,7 +48,7 @@ The harness adapts to the *class* of engine, not the model name:
 | **MEMORY ranking** | relevance-ranked when a local Ollama `embed_model` is reachable, else newest-first | relevance-ranked via `PWN::MemoryIndex` (`~/.pwn/memory.idx`) |
 | **Tool schemas shipped** | all toolsets | `CORE_TOOLS` + top-K keyword matches when `ai.agent.tool_router` is on |
 | **Pre-pass** | none | `plan_first` numbered tool plan before first dispatch |
-| **Intent route** | always | `request_intent` short-circuits how-to (text only) and unauthorized recon (refuse) on all engines; critical for ollama/openwebui |
+| **Intent route** | always | `request_intent` + LLM/heuristic `request_kind` (statement \| question \| autonomous_goal). Short-circuits how-to/questions (text only), greetings/statements (fixed ack), pure recall, and unauthorized recon on all engines; host-evidence Qs (hostname/cwd/whoami) and only true autonomous goals get multi-step TaskSummarizer plans. Critical for ollama/openwebui |
 | **Few-shot** | none | `Learning.exemplars_for(request)` splices a prior successful trace |
 | **Dispatch parsing** | strict | tolerant - Levenshtein tool-name repair + JSON5-ish arg cleanup, each repair fingerprinted into `Mistakes` |
 | **Post-answer** | `auto_introspect` | `auto_introspect` **+** `fact_check_local_final` (auto `extro_verify` on CVE/version-shaped claims) |

data/documentation/Configuration.md

@@ -103,6 +103,8 @@ ai:
     task_summary_every: 5          # When task_summary_verbose: emit Progress every N completed tools.
     task_summary_interval_s: 8.0   # When verbose: also emit when this many seconds elapsed.
     task_summary_verbose: false    # Mid-flight Progress/Finished lines (default: only plan + about_to).
+    task_summary_llm: true         # LLM tangible-task decompose for autonomous goals (default on). false = offline fallback.
+    request_kind_llm: ~            # LLM request_kind classifier (statement|question|autonomous_goal). nil = follow task_summary_llm.
     max_depth: 3                   # Recursion guard: how many levels deep agent_ask/agent_debate sub-agents may spawn sub-agents.
     auto_introspect: true          # Run Learning.auto_introspect (outcome logging + lesson mining) after every final answer.
     auto_extrospect: false         # Optional ambient baseline (host/repo/env ONLY - never launches burpsuite/zaproxy/msf/gqrx). Sense tools (intel/verify/watch/rf_tune/observe) stay on-demand.
@@ -268,6 +270,8 @@ PWN::Config.refresh_env
 | `ai.agent.task_summary_every` | Integer | `5` | `TaskSummarizer.every_n` | Verbose progress cadence (tools). |
 | `ai.agent.task_summary_interval_s` | Float | `8.0` | `TaskSummarizer.interval_s` | Verbose progress cadence (seconds). |
 | `ai.agent.task_summary_verbose` | Boolean | `false` | `TaskSummarizer.verbose?` | Emit mid-flight `Progress:` / `Finished:` lines; default keeps only plan + about_to. |
+| `ai.agent.task_summary_llm` | Boolean \| `nil` | `nil` (on) | `TaskSummarizer.llm_plan_enabled?` | LLM tangible-task decomposition for autonomous goals. `false` forces offline generic fallback (tests / air-gap). |
+| `ai.agent.request_kind_llm` | Boolean \| `nil` | `nil` (follow `task_summary_llm`) | `TaskSummarizer.llm_kind_enabled?` / `request_kind` | LLM classifier for `statement` \| `question` \| `autonomous_goal`. Cheap intents and host-evidence heuristics still win first; `false` is heuristic-only. |
 | `ai.agent.max_depth` | Integer | `3` | `PWN::AI::Agent::Swarm` | Recursion guard for `agent_ask` / `agent_debate` sub-agents spawning sub-agents. |
 | `ai.agent.auto_introspect` | Boolean | `true` | `PWN::AI::Agent::Learning.auto_introspect` | Run outcome logging + lesson mining after every final answer. Toggle live via `learning_auto_introspect_toggle`. |
 | `ai.agent.auto_extrospect` | Boolean | `false` | `PWN::AI::Agent::Extrospection.auto_extrospect` | Optional ambient baseline after every final answer (`AUTO_SECTIONS` = host/repo/env only; never spawns GUI/JVM tools). Prefer on-demand sense tools (`intel`/`verify`/`watch`/`rf_tune`/`observe`). Toggle live via `extro_auto_toggle`. |

data/documentation/How-PWN-Works.md

@@ -28,7 +28,7 @@ hardware).
 | Module | Role |
 |---|---|
 | `Loop` | plan → **TaskSummarizer** briefs → dispatch tool_calls → observe → repeat until final answer; tightens runway when recent turns exhausted the budget |
-| **`TaskSummarizer`** | Executive UX: English tasks primary - `emit_plan!` · `about_to` as `task k/n` + via tools · Loop `plan_context` injection · clearer `plan_idx` advance |
+| **`TaskSummarizer`** | Executive UX: LLM/heuristic `request_kind` gates plans (only autonomous goals, including host-evidence Qs) - English tasks primary - `emit_plan!` · `about_to` as `task k/n` + via tools · Loop `plan_context` injection · clearer `plan_idx` advance |
 | `Registry` | JSON-Schema function definitions grouped into 12 **toolsets** · **82 tools** |
 | `Dispatch` / `Result` | execute a tool, capture stdout/value/error/duration |
 | `PromptBuilder` | inject MEMORY / SKILLS / LEARNING / **KNOWN MISTAKES + FIXES** / METRICS / EXTROSPECTION blocks |

data/documentation/diagrams/dot/task-summarizer.dot

@@ -1,6 +1,6 @@
 digraph "PWN_TaskSummarizer" {
   graph [
-    label=<<B>PWN::AI::Agent::TaskSummarizer - Executive task briefs</B><BR/><FONT POINT-SIZE="11" COLOR="#94a3b8">emit_plan! (full goal once) → about_to (per batch, deduped) → tools on their own lines · Loop single emitter</FONT>>,
+    label=<<B>PWN::AI::Agent::TaskSummarizer - Request kind + executive briefs</B><BR/><FONT POINT-SIZE="11" COLOR="#94a3b8">request_kind → (statement|question: no plan) · (autonomous_goal: emit_plan! → about_to → tools)</FONT>>,
     labelloc=t, fontsize=20, fontname="Helvetica",
     rankdir=TB, splines=spline, nodesep=0.55, ranksep=0.9,
     bgcolor="#0f172a", fontcolor="#e2e8f0", pad=0.6, newrank=true, compound=true
@@ -10,25 +10,34 @@ digraph "PWN_TaskSummarizer" {
   edge [fontname="Helvetica", fontsize=9, color="#94a3b8",
         fontcolor="#cbd5e1", penwidth=1.3, arrowsize=0.8];
 
-  User  [label="👤  User goal", fillcolor="#7dd3fc"];
-  Loop  [label="Loop.run\nts_state = TaskSummarizer.fresh", fillcolor="#c4b5fd"];
-  {rank=same; User; Loop}
+  User  [label="User request", fillcolor="#7dd3fc"];
+  Loop  [label="Loop.run\nts_state = TaskSummarizer.fresh\nrequest_kind + request_intent", fillcolor="#c4b5fd"];
+  Kind  [label="request_kind (LLM + heuristic)\nstatement | question | autonomous_goal\nhost-evidence Qs → goal", fillcolor="#fcd34d", penwidth=2];
+  {rank=same; User; Loop; Kind}
+
+  subgraph cluster_nogoal {
+    label="No multi-step breakdown"; fontcolor="#fda4af";
+    style=rounded; color="#be123c"; bgcolor="#4c0519"; penwidth=2;
+    Stmt [label="statement\nanswer_statement / greeting ack", fillcolor="#fda4af"];
+    Ques [label="question (knowledge/howto/recall)\nanswer_question / howto / recall\nno tools / no plan", fillcolor="#fda4af"];
+  }
+  {rank=same; Stmt; Ques}
 
   subgraph cluster_plan {
-    label="Plan layer (once per turn)"; fontcolor="#ddd6fe";
+    label="Autonomous goal plan (once per turn)"; fontcolor="#ddd6fe";
     style=rounded; color="#6d28d9"; bgcolor="#2e1065"; penwidth=2;
-    Plan   [label="plan(request)\nenumerated steps or heuristic_decompose", fillcolor="#c4b5fd"];
-    EmitP  [label="emit_plan!\nGoal: …\nTangible tasks (N):\n  1. …  2. …", fillcolor="#c4b5fd", penwidth=2];
+    Plan   [label="plan(request)\nenumerated · LLM · fallback\nordered work units", fillcolor="#c4b5fd"];
+    EmitP  [label="emit_plan!\nKind: autonomous_goal\nGoal: ...\nTangible tasks (N):\n  each may use 1+ tools", fillcolor="#c4b5fd", penwidth=2];
   }
   {rank=same; Plan; EmitP}
 
   subgraph cluster_batch {
     label="Batch layer (each tool collection)"; fontcolor="#a7f3d0";
     style=rounded; color="#047857"; bgcolor="#022c22"; penwidth=2;
-    About  [label="about_to(tools:)\ncapability_label\ntool_counts_phrase\nintent_phrase", fillcolor="#6ee7b7"];
+    About  [label="about_to(tools:)\ntask k/n English primary\nvia tools (intent)", fillcolor="#6ee7b7"];
     FP     [label="brief_fingerprint\nlast_brief_fp", fillcolor="#6ee7b7"];
     Dup    [label="duplicate_brief?", shape=diamond, fillcolor="#fcd34d"];
-    EmitB  [label="emit task line\nNext: shell×2 (search)\n[task k/N: …]", fillcolor="#6ee7b7", penwidth=2];
+    EmitB  [label="emit task line\ntask k/N: ...\nvia shell×2 (search)", fillcolor="#6ee7b7", penwidth=2];
     Skip   [label="return nil\n(suppress duplicate)", fillcolor="#fda4af"];
   }
   {rank=same; About; FP; Dup; EmitB; Skip}
@@ -36,7 +45,7 @@ digraph "PWN_TaskSummarizer" {
   subgraph cluster_exec {
     label="Execution (not on task row)"; fontcolor="#fde68a";
     style=rounded; color="#a16207"; bgcolor="#422006"; penwidth=2;
-    Tools  [label="Dispatch tools\nshell · pwn_eval · …\nown REPL lines", fillcolor="#fcd34d"];
+    Tools  [label="Dispatch tools\nshell · pwn_eval · ...\nown REPL lines", fillcolor="#fcd34d"];
     Rec    [label="record!\nadvance plan_idx\n(silent unless verbose)", fillcolor="#fcd34d"];
     Flush  [label="flush! / emit!\noptional closing brief", fillcolor="#fcd34d"];
   }
@@ -45,10 +54,13 @@ digraph "PWN_TaskSummarizer" {
   UI [label="REPL on_tool\nname='task', full text, result=''\n(no truncation)", fillcolor="#7dd3fc"];
 
   User -> Loop [color="#38bdf8", penwidth=2];
-  Loop -> Plan [color="#a78bfa"];
+  Loop -> Kind [color="#fbbf24", penwidth=2];
+  Kind -> Stmt [label="statement", color="#fb7185"];
+  Kind -> Ques [label="question", color="#fb7185"];
+  Kind -> Plan [label="autonomous_goal", color="#a78bfa", penwidth=2];
   Plan -> EmitP [color="#a78bfa"];
   EmitP -> UI [label="full goal once", color="#38bdf8", penwidth=2];
-  Loop -> About [label="before each\ntool batch", color="#a78bfa"];
+  Loop -> About [label="before each\ntool batch\n(goals only)", color="#a78bfa"];
   About -> FP [color="#34d399"];
   FP -> Dup [color="#34d399"];
   Dup -> EmitB [label="new", color="#34d399"];