pwn 0.5.643 → 0.5.650

data/spec/lib/pwn/ai/agent/reward_spec.rb

@@ -35,7 +35,12 @@ describe PWN::AI::Agent::Reward do
     stub_const('PWN::AI::Agent::Reward::DPO_DIR', tmp)
 
     described_class.record_preference(prompt: 'p', rejected: 'bad', chosen: 'good', source: :user_correction)
-    described_class.record_preference(prompt: 'p2', rejected: 'x', chosen: 'y', source: :mistakes_resolve)
+    # P21/P25 — mistakes_resolve requires a trajectory shape
+    described_class.record_preference(
+      prompt: 'p2', rejected: 'failing shell args',
+      chosen: "STRATEGY: x\nWINNING_TRACE:\nshell → uname -r\n#{'t' * 40}",
+      source: :mistakes_resolve, shape: :winning_trace
+    )
     expect(described_class.preferences.length).to eq 2
     expect(described_class.preferences(source: 'user_correction').length).to eq 1
 
@@ -144,4 +149,62 @@ describe PWN::AI::Agent::Reward do
     expect(File.exist?(File.join(tmp, 's.json'))).to be false
     expect(File.exist?(File.join(tmp, 'prefs.jsonl'))).to be true
   end
+  it 'write_source_quota and weak pair geometry (P9)' do
+    tmp = Dir.mktmpdir
+    stub_const('PWN::AI::Agent::Reward::PREFERENCES_FILE', File.join(tmp, 'prefs.jsonl'))
+    r = described_class.record_preference(prompt: 'p', rejected: 'a', chosen: 'CORRECTION: no', source: :critic)
+    expect(r[:skipped]).to eq :weak_pair_geometry
+    11.times do |i|
+      described_class.record_preference(prompt: "p#{i}", rejected: "r#{i}", chosen: "c#{i}", source: :mistakes_resolve, force: true)
+    end
+    q = described_class.write_source_quota(source: 'mistakes_resolve')
+    expect(q[:over_cap]).to be true
+  end
+
+  it 'warm_sentinel fills from learning outcomes (P10)' do
+    tmp = Dir.mktmpdir
+    stub_const('PWN::AI::Agent::Reward::SENTINEL_FILE', File.join(tmp, 's.json'))
+    stub_const('PWN::AI::Agent::Learning::LEARNING_FILE', File.join(tmp, 'l.jsonl'))
+    50.times do |i|
+      PWN::AI::Agent::Learning.note_outcome(task: "t#{i}", success: true, score: 0.8, details: 'd')
+    end
+    r = described_class.warm_sentinel(limit: 60)
+    expect(r[:added]).to be > 0
+  end
+
+  it 'scrub_preferences and usable_preference? (P15)' do
+    tmp = Dir.mktmpdir
+    stub_const('PWN::AI::Agent::Reward::PREFERENCES_FILE', File.join(tmp, 'prefs.jsonl'))
+    described_class.record_preference(
+      prompt: 'p', rejected: 'long rejected text ' * 20,
+      chosen: 'CORRECTION: no', source: :critic, force: true
+    )
+    described_class.record_preference(
+      prompt: 'q', rejected: 'bad',
+      chosen: "WINNING_TRACE:\n#{'shell ok ' * 30}",
+      source: :curriculum, shape: :winning_trace, force: true
+    )
+    dry = described_class.scrub_preferences(dry_run: true)
+    expect(dry[:dropped]).to be >= 1
+    wet = described_class.scrub_preferences(dry_run: false)
+    expect(wet[:after]).to be < wet[:before]
+    bal = described_class.preference_balance(scrub: true)
+    expect(bal).to have_key(:trajectory_fraction)
+  end
+
+  it 'export_dpo reports geometry_dropped (P15)' do
+    tmp = Dir.mktmpdir
+    stub_const('PWN::AI::Agent::Reward::PREFERENCES_FILE', File.join(tmp, 'prefs.jsonl'))
+    stub_const('PWN::AI::Agent::Reward::DPO_DIR', tmp)
+    described_class.record_preference(
+      prompt: 'p', rejected: 'r' * 500, chosen: 'CORRECTION: x', source: :mistakes_resolve, force: true
+    )
+    described_class.record_preference(
+      prompt: 'q', rejected: 'r', chosen: "WINNING_TRACE:\n#{'t' * 100}",
+      source: :counterfactual, shape: :real_dispatch, force: true
+    )
+    info = described_class.export_dpo(out: File.join(tmp, 'd.jsonl'))
+    expect(info[:geometry_dropped]).to be >= 1
+    expect(info[:scrubbed]).to be true
+  end
 end

data/spec/lib/pwn/ai/agent/tools/ruby_eval_spec.rb

@@ -1,10 +1,48 @@
 # frozen_string_literal: true
 
 require 'spec_helper'
+require 'json'
 
 describe 'PWN::AI::Agent::Tools ruby_eval' do
-  it 'registers the pwn_eval tool' do
+  before(:all) do
     PWN::AI::Agent::Registry.discover
+  end
+
+  it 'registers the pwn_eval tool' do
     expect(PWN::AI::Agent::Registry.lookup(name: 'pwn_eval')).not_to be_nil
   end
+
+  it 'returns structured error for SyntaxError payloads (constant block param)' do
+    entry = PWN::AI::Agent::Registry.lookup(name: 'pwn_eval')
+    result = entry.handler.call(code: 'results = probes.map do |R|')
+    expect(result).to be_a(Hash)
+    expect(result[:error]).to match(/SyntaxError/)
+    expect(result[:error]).to match(/\(pwn_eval\)/)
+    expect(result[:error]).not_to match(/ruby_eval\.rb:49/)
+    expect(result[:hint]).to match(/\|r\| not \|R\|/)
+  end
+
+  it 'evaluates valid code and returns stdout + value' do
+    entry = PWN::AI::Agent::Registry.lookup(name: 'pwn_eval')
+    result = entry.handler.call(code: '1 + 1')
+    expect(result[:error]).to be_nil
+    expect(result[:value]).to eq('2')
+  end
+
+  it 'Dispatch.call JSON-wraps syntax errors instead of raising' do
+    out = PWN::AI::Agent::Dispatch.call(
+      tool_call: {
+        id: 't1',
+        type: 'function',
+        function: {
+          name: 'pwn_eval',
+          arguments: JSON.generate(code: 'results = probes.map do |R|')
+        }
+      }
+    )
+    parsed = JSON.parse(out, symbolize_names: true)
+    # handler now rescues → success:true with error key in result
+    expect(parsed[:success]).to eq(true)
+    expect(parsed[:result][:error]).to match(/SyntaxError/)
+  end
 end

data/third_party/pwn_rdoc.jsonl

@@ -13,15 +13,17 @@
 {"messages":[{"role":"user","content":"PWN::AI::Agent::BurpSuite.authors Usage"},{"role":"assistant","content":"`PWN::AI::Agent::BurpSuite.authors`: Author(s)\n\n0day Inc. <support@0dayinc.com>\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::BurpSuite.help Usage"},{"role":"assistant","content":"`PWN::AI::Agent::BurpSuite.help`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.ab_gate Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.ab_gate`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.ab_gate_v2 Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.ab_gate_v2`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.ask_persona Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.ask_persona`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.authors Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.authors`: Author(s)\n\n0day Inc. <support@0dayinc.com>\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.axolotl_config Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.axolotl_config`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.build_eval_set Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.build_eval_set`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.build_telemetry Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.build_telemetry`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.bump_cooldown! Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.bump_cooldown!`: "}]}
-{"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.calibrate Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.calibrate`: Supported Method Parameters\n\nPWN::AI::Agent::Curriculum.calibrate(predicted:, actual:, engine:)\n"}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.calibrate Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.calibrate`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.counterfactual Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.counterfactual`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.critic Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.critic`: Supported Method Parameters\n\nv = PWN::AI::Agent::Curriculum.critic(\n\nrequest: 'required - user request',\nfinal: 'required - candidate final answer',\nsession_id: 'optional - for evidence lookup'\n\n)\n\nReturns { verdict: :pass|:flaw, flaw:, confidence: }. On :flaw the (final, flaw) pair is recorded as a preference so a future self-correction becomes DPO signal.\n"}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.critic_text_only Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.critic_text_only`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.detect_trainer Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.detect_trainer`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.enabled? Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.enabled?`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.ensure_persona Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.ensure_persona`: "}]}
@@ -37,19 +39,26 @@
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.offline_judge Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.offline_judge`: Supported Method Parameters\n\nr = PWN::AI::Agent::Curriculum.offline_judge(\n\nsince_hours: 'optional - lookback window (default 24)',\nlimit: 'optional - max sessions to score (default 40)',\nprm: 'optional - also run Process Reward Model (default true)',\ncommit: 'optional - write scores into learning/sentinel (default true)'\n\n)\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.ollama_create Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.ollama_create`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.practice Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.practice`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.practice_kpi Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.practice_kpi`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.practice_skip? Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.practice_skip?`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.preference_balance Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.preference_balance`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.preference_diet_gate Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.preference_diet_gate`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.python_site_package_roots Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.python_site_package_roots`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.red_team_plan Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.red_team_plan`: Supported Method Parameters\n\nhint = PWN::AI::Agent::Curriculum.red_team_plan(\n\nrequest: 'required - user goal',\nplan: 'required - numbered plan text from plan_first'\n\n)\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.reflect_available? Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.reflect_available?`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.repeating_trend Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.repeating_trend`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.replay_on Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.replay_on`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.replay_on_detailed Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.replay_on_detailed`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.revise_after_flaw Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.revise_after_flaw`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.run_axolotl_trainer Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.run_axolotl_trainer`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.run_trainer Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.run_trainer`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.run_unsloth_trainer Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.run_unsloth_trainer`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.save_cooldown Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.save_cooldown`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.save_models Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.save_models`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.score_branch Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.score_branch`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.score_branch_detailed Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.score_branch_detailed`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.self_play Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.self_play`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.smoke_eval_set Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.smoke_eval_set`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.train_and_gate Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.train_and_gate`: Supported Method Parameters\n\nr = PWN::AI::Agent::Curriculum.train_and_gate(\n\nbase_model: 'optional - ollama base tag (default PWN::Env[:ai][:ollama][:model])',\ntrainer: 'optional - :unsloth | :axolotl | :auto (default :auto)',\ndry_run: 'optional - export + build eval set but do not train (default true)'\n\n)\n\nBest-effort orchestrator. When a supported trainer is installed it produces ~/.pwn/finetune/pwn-vN/, cuts an ollama Modelfile with the LoRA adapter, then REPLAYS Mistakes.top against vN and vN+1 under Reward.judge. Promotes (writes MODELS_FILE) iff vN+1 resolves more signatures. When no trainer is present it still exports SFT+DPO and emits the exact CLI to run manually — so the pipeline is complete even on a box without GPU.\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.trainer_cli_path Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.trainer_cli_path`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Curriculum.trainer_installed? Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Curriculum.trainer_installed?`: "}]}
@@ -251,13 +260,16 @@
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.auto_introspect Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.auto_introspect`: Supported Method Parameters\n\nPWN::AI::Agent::Learning.auto_introspect(\n\nsession_id: 'required - id of the just-completed session',\nrequest: 'optional - original user request (for outcome logging)',\nfinal: 'optional - final assistant answer (for outcome logging)'\n\n)\n\nCalled by Loop.run when PWN::Env[:agent] is truthy. Never raises — learning must not break the primary loop.\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.auto_introspect_enabled? Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.auto_introspect_enabled?`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.build_skill_from_session Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.build_skill_from_session`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.checkable_claim? Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.checkable_claim?`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.compress_exemplar Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.compress_exemplar`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.compress_finetune_trace Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.compress_finetune_trace`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.consolidate Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.consolidate`: Supported Method Parameters\n\nremoved = PWN::AI::Agent::Learning.consolidate(\n\nmax_entries: 'optional - hard cap on PWN::Memory size (default MAX_MEMORY_ENTRIES)'\n\n)\n\nDeduplicates near-identical lesson values and prunes the oldest entries once the cap is exceeded so the injected MEMORY block stays high-signal.\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.distill_skill Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.distill_skill`: Supported Method Parameters\n\nskill = PWN::AI::Agent::Learning.distill_skill(\n\nname: 'required - snake_case name for the new skill',\nsession_id: 'optional - PWN::Sessions id to mine (uses its transcript)',\ncontent: 'optional - explicit markdown body; overrides transcript mining',\nreferences: 'optional - Array of reference URLs / CWE / CVE / ATT&CK ids'\n\n)\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.exemplars_for Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.exemplars_for`: Supported Method Parameters\n\nmsgs = PWN::AI::Agent::Learning.exemplars_for(\n\nrequest: 'required - current user request',\nlimit: 'optional - max exemplar traces to return (default 1)',\nmax_msgs: 'optional - cap on messages per exemplar (default 6)'\n\n)\n\nRetrieval-augmented BEHAVIOUR: keyword-matches request against prior successful outcomes in learning.jsonl, loads the matching session, and compresses its (user, tool, assistant) trace into a short few-shot exemplar Loop.run splices between system and user. Local models are dramatically better with 1 concrete example than with 25 abstract lessons.\n"}]}
-{"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.export_finetune Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.export_finetune`: Supported Method Parameters\n\ninfo = PWN::AI::Agent::Learning.export_finetune(\n\nformat: 'optional - :sharegpt (default) | :openai_jsonl',\nout: 'optional - output path (default ~/.pwn/finetune/pwn-YYYYMMDD.jsonl)',\nmin_tools: 'optional - only sessions with >= N tool messages (default 1)'\n\n)\n\nTurns the learning corpus into a supervised dataset: every session whose learning.jsonl outcome is success:true becomes one training sample (system, user, assistant/tool_calls, tool, …, final). Pair with a weekly PWN::Cron job that runs ‘ollama create <tag>-pwn -f Modelfile` over the export - the only path to ACTUAL parity with a frontier model, because it changes the weights not just the scaffold.\n"}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.export_finetune Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.export_finetune`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.fact_check_local_final Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.fact_check_local_final`: "}]}
-{"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.flip_last_outcome Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.flip_last_outcome`: Supported Method Parameters\n\nPWN::AI::Agent::Learning.flip_last_outcome(\n\nsession_id: 'optional - only flip if the newest outcome belongs to this session',\nreason: 'optional - why it is being flipped (usually the user correction text)'\n\n)\n\nRewrites the most-recently-appended learning.jsonl entry from success:true to success:false. Called by Mistakes.check_user_correction when the user’s next message rejects the previous answer, so the 100 %-success illusion is broken and the failure enters the corpus.\n"}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.flip_last_outcome Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.flip_last_outcome`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.fold_judge_into_metrics Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.fold_judge_into_metrics`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.help Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.help`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.heuristic_lessons Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.heuristic_lessons`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.infer_success Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.infer_success`: "}]}
@@ -266,6 +278,7 @@
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.note_outcome Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.note_outcome`: Supported Method Parameters\n\nentry = PWN::AI::Agent::Learning.note_outcome(\n\ntask: 'required - short description of what was attempted',\nsuccess: 'required - Boolean, did the attempt achieve its goal',\ndetails: 'optional - free-form notes / error / evidence',\nsession_id: 'optional - PWN::Sessions id this outcome belongs to',\ntags: 'optional - Array of String labels for later retrieval'\n\n)\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.outcomes Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.outcomes`: Supported Method Parameters\n\nrows = PWN::AI::Agent::Learning.outcomes(\n\nlimit: 'optional - max entries returned newest-first (default 50)',\nsuccess: 'optional - filter by Boolean outcome',\ntag: 'optional - filter by tag substring'\n\n)\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.purge_noise Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.purge_noise`: Supported Method Parameters\n\nPWN::AI::Agent::Learning.purge_noise\n\nOne-shot GC of the pre-R1 garbage: drops every PWN::Memory entry matching the old ‘SUCCESS: <req> — <final>` / `Avoid repeating failure pattern from <tool>: {“success”:true` shapes. Run once after upgrading; subsequent writes never produce these.\n"}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.recover_predicted_from_session Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.recover_predicted_from_session`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.reflect Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.reflect`: Supported Method Parameters\n\nreport = PWN::AI::Agent::Learning.reflect(\n\nsession_id: 'required - PWN::Sessions id to analyse',\ndry_run: 'optional - when true, do not write to Memory/Skills (default false)'\n\n)\n\nUses PWN::AI::Agent::Reflect (when available) to LLM-summarise the session into structured lessons. Falls back to a heuristic extractor when module_reflection is disabled so learning never stops.\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.reset Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.reset`: Supported Method Parameters\n\nPWN::AI::Agent::Learning.reset\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Learning.semantic_merge Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Learning.semantic_merge`: "}]}
@@ -279,6 +292,7 @@
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Loop.append_session Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Loop.append_session`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Loop.attribute_cause Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Loop.attribute_cause`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Loop.authors Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Loop.authors`: Author(s)\n\n0day Inc. <support@0dayinc.com>\n"}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Loop.budget_exhaustion_hot? Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Loop.budget_exhaustion_hot?`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Loop.calibration_state Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Loop.calibration_state`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Loop.call_engine Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Loop.call_engine`: Supported Method Parameters\n\nmsg = PWN::AI::Agent::Loop.call_engine(\n\nmessages: 'required - OpenAI-format messages array',\ntools: 'optional - OpenAI tools array'\n\n)\n\nReturns a normalised assistant message hash:\n\n{ role: 'assistant', content: String|nil,\n  tool_calls: [ {id:, type:'function', function:{name:, arguments:}} ],\n  _native_content: <provider raw>  (when adapter needs round-trip) }\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Loop.compact_history! Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Loop.compact_history!`: "}]}
@@ -302,12 +316,19 @@
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.calibration Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.calibration`: Supported Method Parameters\n\ncal = PWN::AI::Agent::Metrics.calibration(engine: :ollama)\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.changepoints Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.changepoints`: Supported Method Parameters\n\ncps = PWN::AI::Agent::Metrics.changepoints\n\nE1 — tools whose CUSUM tripped (success_rate regime change). The caller (Mistakes.record / Curriculum) triggers extro_snapshot + correlate on these so a Mistake caused by env drift is tagged cause: :env_drift and does NOT count toward [REPEATING].\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.decay Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.decay`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.effective_rate Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.effective_rate`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.gamma_sample Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.gamma_sample`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.help Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.help`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.judge_confidence Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.judge_confidence`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.judge_rate Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.judge_rate`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.load Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.load`: Supported Method Parameters\n\nmetrics = PWN::AI::Agent::Metrics.load\n"}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.prm_advantage Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.prm_advantage`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.prm_n Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.prm_n`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.proxy_trust Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.proxy_trust`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.record Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.record`: Supported Method Parameters\n\nPWN::AI::Agent::Metrics.record(\n\nname: 'required - tool name that was dispatched',\nsuccess: 'required - Boolean, did the handler complete without error',\nduration: 'optional - Float seconds the dispatch took',\nerror: 'optional - String error message when success is false',\nengine: 'optional - Symbol/String AI engine that chose this tool (segments telemetry)'\n\n)\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.record_calibration Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.record_calibration`: Supported Method Parameters\n\nPWN::AI::Agent::Metrics.record_calibration(predicted:, actual:, brier:, engine:)\n\nW3 — plan_first emits p(success); Loop.run calls this with the realised outcome. Tracked per-engine so calibration of the local LoRA vs frontier is comparable.\n"}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.record_judge Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.record_judge`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.record_step_reward Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.record_step_reward`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.reset Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.reset`: Supported Method Parameters\n\nPWN::AI::Agent::Metrics.reset\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.save Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.save`: Supported Method Parameters\n\nPWN::AI::Agent::Metrics.save(\n\nmetrics: 'required - Hash returned by .load / mutated in place'\n\n)\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Metrics.summary Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Metrics.summary`: Supported Method Parameters\n\nrows = PWN::AI::Agent::Metrics.summary(\n\nlimit: 'optional - cap number of tools returned (default 25)',\nengine: 'optional - only that engine\\'s sub-bucket (falls back to global when absent)'\n\n)\n"}]}
@@ -382,30 +403,38 @@
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.empty_sentinel Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.empty_sentinel`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.export_dpo Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.export_dpo`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.extract_cmd Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.extract_cmd`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.fold_step_reward_to_metrics Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.fold_step_reward_to_metrics`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.generator_mix Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.generator_mix`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.help Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.help`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.heuristic_judge Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.heuristic_judge`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.heuristic_prm Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.heuristic_prm`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.infer_shape Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.infer_shape`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.judge Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.judge`: Supported Method Parameters\n\nv = PWN::AI::Agent::Reward.judge(\n\nrequest: 'required - original user request',\nfinal: 'required - assistant final answer',\nsession_id: 'optional - PWN::Sessions id (adds tool trace)',\ntrace: 'optional - Array of tool-result strings (overrides session_id)',\ncommit: 'optional - write score into learning.jsonl / sentinel (default true)'\n\n)\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.llm_judge Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.llm_judge`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.llm_prm Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.llm_prm`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.load_sentinel Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.load_sentinel`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.load_trace Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.load_trace`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.normalize_sentinel Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.normalize_sentinel`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.preference_balance Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.preference_balance`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.preferences Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.preferences`: Supported Method Parameters\n\nrows = PWN::AI::Agent::Reward.preferences(limit: 500, source: nil)\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.prm Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.prm`: Supported Method Parameters\n\nsteps = PWN::AI::Agent::Reward.prm(\n\nrequest: 'required - user goal',\nsession_id: 'optional - session to score in place',\ntrace: 'optional - Array of {name:, args:, result:} or Strings'\n\n)\n\nReturns [{idx:, step:, reward: -1|0|1}, …] and, when session_id is given, rewrites each tool line in the transcript with a ‘[step_reward=N]` prefix so exemplars_for / distill_skill can keep only reward>0 steps (C4 minimal sufficient trace).\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.proxy_distrust Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.proxy_distrust`: "}]}
-{"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.record_preference Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.record_preference`: Supported Method Parameters\n\nPWN::AI::Agent::Reward.record_preference(\n\nprompt: 'required - the context / user request',\nrejected: 'required - the losing completion / action',\nchosen: 'required - the winning completion / action',\nsource: 'optional - :user_correction | :mistakes_resolve | :counterfactual | :critic'\n\n)\n"}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.record_preference Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.record_preference`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.record_sentinel Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.record_sentinel`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.recoverable_shape Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.recoverable_shape`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.reflect_available? Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.reflect_available?`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.reset Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.reset`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.reset_sentinel Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.reset_sentinel`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.scrub_preferences Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.scrub_preferences`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.semantic_ok Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.semantic_ok`: Supported Method Parameters\n\nh = PWN::AI::Agent::Reward.semantic_ok(\n\nname: 'required - tool name',\nraw: 'required - JSON string returned by Dispatch.call',\nargs: 'optional - the tool call arguments (used for BENIGN_EXIT)'\n\n)\n\nReturns { ok:, semantic_ok:, exit:, err:, benign: }. :ok is the old proxy (handler didn’t raise); :semantic_ok additionally knows that grep/diff/find exit≠0 with empty stderr is not a failure. Loop.run records Metrics on :ok but only records Mistakes on !semantic_ok.\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.sentinel Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.sentinel`: Supported Method Parameters\n\nr = PWN::AI::Agent::Reward.sentinel\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.set_proxy_distrust Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.set_proxy_distrust`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.usable_preference? Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.usable_preference?`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.user_correction_rate Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.user_correction_rate`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.verify_as_reward Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.verify_as_reward`: Supported Method Parameters\n\ng = PWN::AI::Agent::Reward.verify_as_reward(final: text)\n"}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.warm_sentinel Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.warm_sentinel`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.window_means Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.window_means`: "}]}
+{"messages":[{"role":"user","content":"PWN::AI::Agent::Reward.write_source_quota Usage"},{"role":"assistant","content":"`PWN::AI::Agent::Reward.write_source_quota`: "}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::SAST.analyze Usage"},{"role":"assistant","content":"`PWN::AI::Agent::SAST.analyze`: Supported Method Parameters\n\nai_analysis = PWN::AI::Agent::SAST.analyze(\n\nrequest: 'required - A string containing the source code snippet to be analyzed for SAST antipatterns and vulnerabilities.'\n\n)\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::SAST.authors Usage"},{"role":"assistant","content":"`PWN::AI::Agent::SAST.authors`: Author(s)\n\n0day Inc. <support@0dayinc.com>\n"}]}
 {"messages":[{"role":"user","content":"PWN::AI::Agent::SAST.help Usage"},{"role":"assistant","content":"`PWN::AI::Agent::SAST.help`: "}]}

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: pwn
 version: !ruby/object:Gem::Version
-  version: 0.5.643
+  version: 0.5.650
 platform: ruby
 authors:
 - 0day Inc.