Fable-mode

Overview Schema Source Related Servers Discussions Admin

Fable session, evidence, and review workflow

fable_session

Destructive

Run a Fable operation for gated AI coding workflows: create sessions, record evidence, verify proofs, advance phases, and control write access.

Instructions

Runs one Fable operation selected by action. Use it for Fable sessions, gates, evidence/proofs, review rubrics, checkpoints, compression, design audits, research scrapers, and explicitly experimental System 3 operations; use browser_* tools for web-page navigation and interaction. Start stateful workflows with create_session and a session_name; later session actions reuse that name. Only fields documented for the selected action are read, and missing or invalid inputs return an Error: message without raising an MCP transport error. Status/list/telemetry/get/view/check operations are read-only. Create, log, record, set, advance, checkpoint, restore, track, evaluate, register, compress/accumulate, evolve, generate, and scraper operations can persist Fable state or artifacts. unlock_execution remains blocked until its time and rationale gates pass; apply_auto_update can replace installed Fable files. The result is action-specific human-readable text in result; inspect that text for success or Error: before continuing.

Input Schema

TableJSON Schema

Name	Required	Description
`tag`	No	Epistemic classification tag for truth calibration.
`code`	No	Source code content (HTML, JSX, CSS, or TSX) to audit or validate against anti-slop gates.
`name`	No	Name identifier for automation pipeline spec.
`tree`	No	Tree hierarchy adjacency list or nested dict for hyperbolic embedding.
`claim`	No	Fact, hypothesis statement, or unknown parameter to track in epistemic ledger.
`edges`	No	List of directed edge dictionaries for Causal DAG construction.
`gamma`	No	Policy precision inverse temperature gamma for Active Inference softmax (default 16.0).
`label`	No	Optional label or description for compressed CAS node.
`nodes`	No	List of node dictionaries for Causal DAG construction.
`action`	Yes	The Fable session action to perform.
`axioms`	No	List of reference axiom names or strings for proof oracle.
`domain`	No	Domain boundary for the invariant or cortical plasticity consolidation.
`states`	No	List of hidden state identifiers for Active Inference POMDP.
`target`	No	Target URL, ID, subreddit, repo, paper ID, handle, or search query string for research scraping actions.
`worlds`	No	List of world/state definitions for Kripke structure.
`actions`	No	List of control action identifiers for Active Inference POMDP.
`cas_ref`	No	CAS reference URI (cas://<sha256_hex>) or 64-char hash.
`content`	No	Raw text content or payload to store/compress into Content-Addressed Storage (CAS).
`context`	No	Hypothesis typing context dictionary {name: type} for proof oracle.
`d_prior`	No	Prior initial state belief distribution D for Active Inference.
`formula`	No	CTL / modal formula string to verify against Kripke structure (e.g. 'AG(safe)', 'EF(goal)').
`mockups`	No	List of concept dictionaries for visual mockup recording.
`payload`	No	Micro-payload text for adaptive batch accumulator or CAS storage.
`root_id`	No	Root node ID for hyperbolic tree embedding or initial world.
`task_id`	No	Task identifier for cortical plasticity consolidation or session lineage.
`a_matrix`	No	Observation likelihood matrix A [O x S] for Active Inference.
`criteria`	No	List of criteria pointers or JSON string of rubric items for goal score evaluation.
`end_line`	No	Ending line number (1-indexed inclusive) for windowed line slice viewing.
`evidence`	No	Source file, command output, line number, or URL supporting the claim.
`metadata`	No	Optional metadata dictionary attached to accumulated micro-payload.
`policies`	No	List of candidate action sequence policies for Expected Free Energy evaluation.
`archetype`	No	Haute aesthetic archetype identifier for design tokens or scaffolding.
`base_step`	No	Base geodesic step distance for hyperbolic tree embedding (default 1.0).
`curvature`	No	Sectional curvature c > 0 of Poincaré manifold (default 1.0).
`dimension`	No	Dimension of Poincaré ball manifold (default 2).
`file_path`	No	Target file path for file change tracking or proof verification.
`objective`	No	High-level goal or problem statement for the reasoning session.
`rationale`	No	Justification for unlocking code execution after satisfying cognitive gates.
`rubric_id`	No	Identifier of the goal rubric.
`b_matrices`	No	State transition matrices B [Action -> S x S] for Active Inference.
`entrypoint`	No	Optional function name to probe inside the target source for red-team review. Auto-detected when the source defines exactly one public function.
`focus_area`	No	Specific subsystem, component, algorithm, or interface being critically re-evaluated.
`model_name`	No	Name for the causal DAG model in System 3 simulation.
`next_phase`	No	Target phase to advance the session into.
`proof_type`	No	Deterministic proof type.
`start_line`	No	Starting line number (1-indexed inclusive) for windowed line slice viewing.
`change_type`	No	Classification of file change.
`evaluations`	No	Alias for item_evaluations: list or dict of criterion evaluation updates.
`force_flush`	No	Force flush buffered micro-payloads immediately into a composite frame.
`generations`	No	Number of evolutionary generations to run (default 3).
`node_labels`	No	Optional mapping of node IDs to readable labels.
`observation`	No	Current sensory observation for Active Inference belief updating.
`target_code`	No	Source-code string input. Loaded and probed in an isolated subprocess sandbox (process boundary, resource limits, per-call timeout).
`target_name`	No	Identifier or name of the target module, class, or function for red-team review.
`transitions`	No	List of transitions or adjacency dictionary for Kripke structure.
`user_prompt`	No	Alias for prompt: high-level design prompt describing desired web experience.
`code_snippet`	No	Alternative alias for source-code string input. Executed in the isolated subprocess sandbox.
`diff_summary`	No	Concise summary of file changes, diff, or slated edits.
`observations`	No	List of observation identifiers for Active Inference POMDP.
`prior_report`	No	Prior red-team breakage report dictionary or JSON string to verify remediation against.
`session_name`	No	Unique identifier / name for the Fable session.
`target_score`	No	Target composite goal score threshold (default 0.95, min 0.0, max 1.0).
`thesis_title`	No	Title of the thesis paradigm for System 3 dialectical synthesis.
`artifact_path`	No	Absolute filesystem path to blueprint artifact, proof, or benchmark script documenting the refinement.
`c_preferences`	No	Prior preference distribution C over observations for Active Inference.
`evaluator_cmd`	No	Alias for evaluator_command: shell command or tool invocation for candidate evaluator.
`failure_count`	No	Historical failure count for tri-level cognitive gear arbitration.
`failure_modes`	No	List of adversarial failure modes identified in critique.
`generator_cmd`	No	Alias for generator_command: shell command or tool invocation for candidate generator.
`initial_world`	No	Initial world ID for Kripke model checking.
`interventions`	No	Dictionary of Pearl's do-operator interventions: {node_id: value}.
`mutation_rate`	No	Genetic mutation rate probability (default 0.15).
`phase_summary`	No	Concise summary of findings or deliverables completed in the previous phase.
`pipeline_name`	No	Alternative alias for automation pipeline name.
`pipeline_type`	No	Type of autonomous pipeline (default 'closed_loop').
`target_metric`	No	Target KPI node ID for sensitivity and structural brittleness analysis.
`contradictions`	No	List of parameter trade-offs / contradictions to resolve with TRIZ principles.
`crossover_rate`	No	Genetic crossover rate probability (default 0.80).
`dials_override`	No	Optional discrete dials override: variance (1-10), motion (1-10), density (1-10).
`invariant_name`	No	Identifier or title of the formal invariant (e.g., 'INV-01: Zero-Deadlock Ring Buffer').
`max_iterations`	No	Maximum closed-loop iterations before halting (default 10).
`seed_paradigms`	No	Optional list of initial paradigm definitions to seed the gene pool.
`task_objective`	No	Task objective or target outcome for goal scoring rubric.
`min_improvement`	No	Minimum scalar measured-fitness gain required to reset stagnation (default 0.001).
`population_size`	No	Population size for evolutionary gene pool (default 12).
`refinement_type`	No	Type of rethink-refine cycle (e.g. 'archetype_exploration', 'triz_resolution', 'adversarial_falsification', 'benchmark_probe', 'failure_mode_analysis').
`remediated_code`	No	Remediated source-code string input. Executed in the isolated subprocess sandbox.
`subagent_prompt`	No	The delegation prompt or contract text for the subagent to validate.
`target_resource`	No	Target resource (file path, receipt ID) for proof verification.
`task_complexity`	No	Task complexity index [0.0, 1.0] for tri-level cognitive arbitration.
`antithesis_title`	No	Title of the antithesis / adversarial critique.
`broken_scenarios`	No	List or JSON string of broken attack scenarios found during red-team review.
`formal_statement`	No	Mathematical or formal contract specification that must never be violated.
`item_evaluations`	No	Item evaluations mapping, list, or JSON string with pointer_id, satisfied, score, evidence_receipt_id.
`observed_fitness`	No	Optional measured fitness object keyed by genome ID, with '*' as a default row. Supplying it enables bounded evidence-fed evolution with stagnation stopping; values are clamped to [0, 1] and never authorize deployment.
`selected_concept`	No	Identifier of the selected visual concept archetype.
`target_threshold`	No	Target threshold score for pipeline iteration termination (default 0.95).
`custom_hypotheses`	No	List or JSON string of custom adversarial hypotheses / attack vectors.
`evaluator_command`	No	Shell command or tool invocation for candidate evaluator.
`generator_command`	No	Shell command or tool invocation for candidate generator.
`max_debate_rounds`	No	Maximum number of dialectical debate rounds for synthesis (default 4).
`objective_weights`	No	Optional dictionary of weights across the 10 Pareto dimensions.
`archetype_override`	No	Optional override for Haute aesthetic archetype.
`auto_log_epistemic`	No	Whether to automatically record retrieved research findings as a [HYPOTHESIS] candidate item in Fable Session's epistemic ledger for subsequent cross-verification.
`proof_or_rationale`	No	Proof sketch, rationale, or inductive argument establishing the invariant.
`thesis_description`	No	Core architecture description and assumptions for the thesis candidate.
`affected_invariants`	No	List or string of invariant names affected by this file change.
`stagnation_patience`	No	Measured generations without improvement before automatic evolution stops (default min(4, generations)).
`time_budget_minutes`	No	Immutable outer authority budget in minutes. set_timer can only change the internal pacing timer.
`include_line_numbers`	No	Whether to format line slice with line numbers.
`contradiction_density`	No	Contradiction density index [0.0, 1.0] for cognitive gear shifting.
`epistemic_uncertainty`	No	Epistemic uncertainty index [0.0, 1.0] for arbitration.
`critique_or_bottleneck`	No	Critical flaw, vulnerability, edge case, memory/latency bottleneck, or assumption scrutinized.
`terminal_probe_results`	No	Live empirical probe output, benchmark figures, latency numbers, or profiling stats supporting the refinement.
`architectural_refinement`	No	Concrete architectural evolution, optimization, or algorithm change derived from the critique.
`target_residual_threshold`	No	Target residual contradiction score threshold for convergence (default 0.15).

Output Schema

TableJSON Schema

Name	Required	Description	Default
`result`	Yes	Action-specific result text. Failures begin with `Error:`.

Schema Changelog

Changes observed during successful MCP inspections.

Changed3 schema fields changedv1.3.9Sep 19, 2026

addedInput schema / properties / min_improvement

Added value: +{
+  "description": "Minimum scalar measured-fitness gain required to reset stagnation (default 0.001).",
+  "minimum": 0,
+  "type": "number"
+}

addedInput schema / properties / observed_fitness

Added value: +{
+  "description": "Optional measured fitness object keyed by genome ID, with '*' as a default row. Supplying it enables bounded evidence-fed evolution with stagnation stopping; values are clamped to [0, 1] and never authorize deployment.",
+  "type": [
+    "object",
+    "string"
+  ]
+}

addedInput schema / properties / stagnation_patience

Added value: +{
+  "description": "Measured generations without improvement before automatic evolution stops (default min(4, generations)).",
+  "minimum": 1,
+  "type": "integer"
+}

Changed5 schema fields changedv1.3.7Sep 17, 2026

changedInput schema / properties / action / enum

Previous value: -[
-  "create_session",
-  "set_timer",
-  "get_status",
-  "telemetry",
-  "advance_phase",
-  "log_epistemic_item",
-  "record_invariant",
-  "log_refinement_cycle",
-  "unlock_execution",
-  "checkpoint_session",
-  "restore_session",
-  "list_sessions",
-  "compile_delegation_contract",
-  "compress_payload",
-  "decompress_payload",
-  "view_slice",
-  "accumulate_payload",
-  "flush_accumulator",
-  "get_compression_stats",
-  "system3_dialectical_synthesis",
-  "system3_causal_simulate",
-  "system3_evolve_paradigms",
-  "system3_induce_axioms",
-  "system3_meta_reflect",
-  "system3_tri_level_orchestrate",
-  "system3_hyperbolic_embed",
-  "system3_kripke_verify",
-  "system3_active_inference",
-  "system3_proof_oracle",
-  "track_file_change",
-  "get_session_lineage",
-  "inspect_plan",
-  "verify_proof",
-  "record_visual_mockups",
-  "validate_event_history",
-  "set_goal_rubric",
-  "evaluate_goal_rubric",
-  "get_goal_rubric",
-  "register_automation_pipeline",
-  "red_team_code_review",
-  "record_breakage_report",
-  "verify_red_team_remediation",
-  "cortical_define_lobe",
-  "cortical_list_lobes",
-  "check_auto_update",
-  "apply_auto_update",
-  "evolve_cortex",
-  "audit_anti_slop",
-  "infer_design_brief",
-  "generate_design_tokens",
-  "generate_awwwards_scaffold",
-  "validate_preflight_design",
-  "list_design_archetypes",
-  "scrape_web",
-  "scrape_youtube",
-  "scrape_reddit",
-  "scrape_x",
-  "scrape_github",
-  "scrape_arxiv"
-]New value: +[
+  "create_session",
+  "set_timer",
+  "get_status",
+  "telemetry",
+  "advance_phase",
+  "log_epistemic_item",
+  "record_invariant",
+  "log_refinement_cycle",
+  "unlock_execution",
+  "checkpoint_session",
+  "restore_session",
+  "list_sessions",
+  "compile_delegation_contract",
+  "compress_payload",
+  "decompress_payload",
+  "view_slice",
+  "accumulate_payload",
+  "flush_accumulator",
+  "get_compression_stats",
+  "system3_dialectical_synthesis",
+  "system3_causal_simulate",
+  "system3_evolve_paradigms",
+  "system3_induce_axioms",
+  "system3_meta_reflect",
+  "system3_tri_level_orchestrate",
+  "system3_hyperbolic_embed",
+  "system3_kripke_verify",
+  "system3_active_inference",
+  "system3_proof_oracle",
+  "track_file_change",
+  "get_session_lineage",
+  "inspect_plan",
+  "verify_proof",
+  "adjudicate_evidence",
+  "record_visual_mockups",
+  "validate_event_history",
+  "set_goal_rubric",
+  "evaluate_goal_rubric",
+  "get_goal_rubric",
+  "register_automation_pipeline",
+  "red_team_code_review",
+  "record_breakage_report",
+  "verify_red_team_remediation",
+  "cortical_define_lobe",
+  "cortical_list_lobes",
+  "check_auto_update",
+  "apply_auto_update",
+  "evolve_cortex",
+  "audit_anti_slop",
+  "infer_design_brief",
+  "generate_design_tokens",
+  "generate_awwwards_scaffold",
+  "validate_preflight_design",
+  "list_design_archetypes",
+  "scrape_web",
+  "scrape_youtube",
+  "scrape_reddit",
+  "scrape_x",
+  "scrape_github",
+  "scrape_arxiv"
+]

changedInput schema / properties / code_snippet / description

Previous value: -"Alternative alias for source-code string input. Dynamic target execution from source-code strings is disabled and requires a separate sandboxed executor."New value: +"Alternative alias for source-code string input. Executed in the isolated subprocess sandbox."

addedInput schema / properties / entrypoint

Added value: +{
+  "description": "Optional function name to probe inside the target source for red-team review. Auto-detected when the source defines exactly one public function.",
+  "type": "string"
+}

changedInput schema / properties / remediated_code / description

Previous value: -"Remediated source-code string input. Dynamic target execution from source-code strings is disabled and requires a separate sandboxed executor."New value: +"Remediated source-code string input. Executed in the isolated subprocess sandbox."

changedInput schema / properties / target_code / description

Previous value: -"Source-code string input. Dynamic target execution from source-code strings is disabled and requires a separate sandboxed executor."New value: +"Source-code string input. Loaded and probed in an isolated subprocess sandbox (process boundary, resource limits, per-call timeout)."

Changed1 schema field changedv1.3.5Sep 16, 2026

changedOutput schema / (root)

Previous value: -nullNew value: +{
+  "additionalProperties": false,
+  "properties": {
+    "result": {
+      "description": "Action-specific result text. Failures begin with `Error:`.",
+      "type": "string"
+    }
+  },
+  "required": [
+    "result"
+  ],
+  "type": "object"
+}

First observedv1.3.2Sep 15, 2026

TDQS

A4.8/5.0

Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations declare destructiveHint=true and readOnlyHint=false, and the description goes well beyond them: it partitions operations into read-only families (status/list/telemetry/get/view/check) vs. stateful ones (create, log, record, checkpoint, etc.), discloses error semantics ('missing or invalid inputs return an `Error:` message without raising an MCP transport error'), warns that `unlock_execution` stays blocked until gates pass, and flags that `apply_auto_update` can replace installed Fable files. This materially enriches the annotation-level safety profile.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single dense paragraph of moderate length, but for a tool with 116 parameters and 61 actions, the density is justified and nearly every clause earns its place. It is front-loaded with the core dispatcher concept and action-selection rule. Minor deduction: the long comma-enumerated lists of operation names and families are harder to scan than structured bullets or a small table would be.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the extreme complexity (116 params, 61 actions, output schema present), the description covers the essentials an agent needs: dispatch semantics, sibling-tool routing, session lifecycle, read-only vs. persistent operations, error-handling behavior, and two dangerous operations that require caution. Because an output schema exists, not restating return-value structure is acceptable; nothing critical is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the schema already documents every parameter, giving a baseline of 3. The description adds genuinely useful parameter behavior beyond the schema: 'Only fields documented for the selected action are read' — a critical clarification for a 116-parameter dispatcher — plus the `session_name` reuse convention and the note that results live in `result`. These rules prevent an agent from assuming all 116 parameters are applicable to every call.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description opens with a specific verb and resource: 'Runs one Fable operation selected by `action`' — accurately framing this as an action-dispatching umbrella tool. It then enumerates the domains it covers (sessions, gates, evidence, rubrics, checkpoints, etc.) and explicitly distinguishes itself from the browser_* sibling family, so an agent can determine what this tool is versus what it is not.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Provides explicit when-to-use guidance ('Use it for Fable sessions, gates, evidence/proofs...') and explicit when-not-to-use routing ('use browser_* tools for web-page navigation and interaction'). It also gives a concrete workflow rule: start stateful flows with `create_session` and `session_name`, then reuse the name — actionable direction an agent can follow without opening the schema.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server