Atlas / MCP servers / juyterman1000 / Entroly

EntrolyBLOCK

Name: Entroly
Author: juyterman1000

mcp/juyterman1000/entroly

MCP server by juyterman1000 · agent · ai agents · ai coding agents · Audited 2026-10-01

Cut AI context cost without trusting the compressor. Every reduction is reversible, byte-exact recoverable, and carries an auditable receipt. Local-first, works through proxy, MCP, SDK, or agent wrapper.

Verdict

BLOCK

Grade

Trust score

41 /100

Exposed tools

114 89r · 22w · 3d

Transport

sse · stdio

License

Apache-2.0

Stars

471

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

Entroly — Select AI context with auditable recovery.

Receipt-backed selection records: what was kept, what was omitted, and the handle that recovers the exact original bytes. Compression you can undo, on your own repository, in one command — without replacing your model or agent architecture.

code --install-extension entroly.entroly-vscode · pip install -U entroly && entroly go · npx entroly

Entroly is an open-source, local-first AI token-efficiency and Context Assurance layer: budgeted evidence selection, recoverable context compression, content-addressed evidence recovery, and auditable receipts. Works through VS Code, Claude Code, Cursor, Codex, OpenClaw, GitHub Copilot, Aider, and OpenAI/Anthropic-compatible apps.

Read from source at commit 946959a19d7fOBSERVED · 2026-10-01

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.

claude-code (pypi)

claude mcp add entroly -- uvx entroly==1.0.85

claude-code (npm)

claude mcp add entroly-mcp -- npx -y [email protected]

Exposed tools (114)

89 read · 22 write · 3 destructive. Blast radius: 3 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

Tool	Risk	Description
`advance_proof_guided_context`	read	Verify one model round and return exact evidence or a final answer.
`allowed`	read	return value + 1
`analyze_codebase_health`	read	Analyze codebase health (grade A-F).
`blast_radius`	read	Analyze the blast radius of file changes on existing beliefs. Returns affected beliefs, risk level, and recommendations.
`checkpoint_state`	write	Save current state for crash recovery.
`compile_beliefs`	read	Compile source code into belief artifacts (Truth → Belief pipeline). Scans directory for source files, extracts entities, resolves dependencies, writes beliefs.
`compile_docs`	read	Compile markdown documentation files into belief artifacts with confidence 0.80 (human-authored > machine-inferred).
`compress_image`	read	Compress an image for vision model input, with optional OCR text extraction.
`compress_shell`	read	Compress CLI output (git, npm, cargo, pytest, etc.) preserving errors and key info.
`coverage_gaps`	read	Find source files with no corresponding belief in the vault. Identifies blind spots before running compile_beliefs.
`create_context_receipt`	write	Create a Context Receipt from supplied documents.
`create_context_receipt_from_path`	write	Create a Context Receipt from a local document file or directory.
`create_skill`	write	Create a new skill from a capability gap (Evolution layer). Generates SKILL.md, tool.py, metrics.json, and test cases.
`denied`	read	return
`eicv_suppress_hallucinations`	read	Verify an LLM response and optionally rewrite hallucinated claims.
`eicv_verify_claim`	read	Verify a single claim against evidence using the EICV pipeline.
`entroly-context`	write	Show Entroly context savings or run
`entroly_dashboard`	read	Show live value metrics: money saved, performance, bloat prevention, quality.
`entroly_retrieve`	read	Retrieve exact source content omitted by compressed context.
`epistemic_route`	read	Route a query through the CogOps Epistemic Ingress Controller. Inspects 4 signals (intent, belief coverage, freshness, risk) and selects one of 5 canonical flows.
`execute_flow`	write	Execute a full canonical epistemic flow end-to-end. Routes query through the Ingress Controller then chains the appropriate pipeline.
`explain_context`	read	Explain why each fragment was included/excluded in last optimization.
`explain_receipt_omission`	read	Explain why a chunk was omitted from a Context Receipt.
`export_training_data`	read	Export vault beliefs as JSONL training data for LLM finetuning. Filters out stale and low-confidence beliefs.
`get_stats`	read	Get comprehensive session statistics.
`ingest_diagram`	read	Ingest an image/diagram into context memory (requires PIL/OCR — not available in WASM runtime, use pip install entroly).
`ingest_diff`	write	Ingest a code diff/patch into context memory. Converts unified diff into structured change summary with intent classification.
`ingest_voice`	read	Ingest audio into context memory (requires whisper — not available in WASM runtime, use pip install entroly).
`inspect_proof_guided_context`	read	Inspect the last durable proof-guided response without advancing it.
`learn_from_failures`	write	Mine failure patterns from PRISM, vault, and evolution data. Optionally apply corrections.
`list_compression_receipts`	read	List only receipts visible to this workspace/session scope.
`locate_evidence`	read	Locate exact source spans without generating an answer.
`manage_skills`	read	Manage CogOps skill lifecycle (Evolution layer). Actions: list, benchmark, promote.
`optimize_context`	read	Select the optimal context subset for a token budget. Uses IOS + PRISM RL + Channel Coding.
`prefetch_related`	read	Predict which files the LLM will need next based on co-access patterns and dependency graph.
`prepare_proof_guided_context`	read	Prepare a durable proof-guided model request from local documents.
`prepare_task_dream`	read	Prepare an expiring, receipt-backed task skill before agent work.
`process_change`	read	Process a code change through the Change-Driven pipeline (Flow 4). Diff → ChangeSet → Review → Blast Radius → Vault.
`recall_relevant`	read	Semantic recall of most relevant stored fragments.
`record_command_exit`	read	Record the exit code of a command that was generated and executed.
`record_edit_outcome`	write	Record whether the user accepted, reverted, or retried an AI edit.
`record_outcome`	read	Record whether selected fragments led to success/failure (RL feedback).
`record_test_result`	read	Record that tests RAN and either passed or failed for a request.
`recover_receipt_omission`	read	Recover the full text of context a Context Receipt omitted.
`recover_shell`	read	Recover full CLI output from a compression handle.
`refresh_beliefs`	read	Mark beliefs as stale after file changes so the next verify pass will flag them for re-compilation.
`refresh_repository_index`	read	Atomically rebuild the fixed repository snapshot.
`remember_fragment`	read	Store a context fragment with automatic dedup and entropy scoring.
`render_context_receipt`	read	Render a Context Receipt JSON artifact as a Markdown audit report.
`repo_file_map`	read	Return the canonical Entroly file map across Python, Rust core, and WASM repos with ownership roles.
`repository_architecture`	read	Return verified layers, communities, cycles, routes, and hotspots.
`repository_architecture_diff`	read	Compare two committed architecture receipts and report structural drift.
`repository_change_impact`	read	Return reverse file/call impact for known workspace-relative paths.
`repository_code_health`	read	Audit verified complexity, cycles, coupling, and navigability risk.
`repository_file_move_apply`	write	Apply a blocker-free module move after commitment and risk checks.
`repository_file_move_preview`	write	Preview a headless Python module move with exact import rewrites.
`repository_git_architecture_diff`	write	Compare a local Git commit with the verified worktree without checkout.
`repository_graph_query`	read	Query verified typed neighbors, paths, relatedness, or impact.
`repository_graph_snapshot`	read	Export the complete bounded graph as a portable committed snapshot.
`repository_graph_snapshot_check`	read	Verify a shared snapshot and prove whether it matches current sources.
`repository_http_routes`	read	Discover verified HTTP routes, mounted prefixes, handlers, and collisions.
`repository_interprocedural_flow`	read	Trace source-verified arguments, parameters, returns, and consumers.
`repository_lsp_rename_preview`	write	Run the operator-configured LSP and return a committed no-write plan.
`repository_map`	read	Rank a receipt-backed structural map across the fixed repository.
`repository_program_graph`	read	Return verified Python control flow and reaching definitions.
`repository_program_slice`	read	Build a proof-carrying partial code slice from verified facts.
`repository_rename_apply`	write	Apply a previewed rename after plan-hash and risk acknowledgement.
`repository_rename_preview`	write	Preview exact rename edits; performs no writes and reports incompleteness.
`repository_runtime_overlay`	read	Bind value-free runtime events to fresh source and symbol evidence.
`repository_safe_delete_apply`	destructive	Apply a blocker-free delete after plan-hash and risk acknowledgement.
`repository_safe_delete_preview`	destructive	Preview a headless delete; any known or lexical reference blocks it.
`repository_semantic_overlay`	read	Verify external LSP/compiler ranges before trusting semantic edges.
`repository_summary`	read	Return bounded counts and the deterministic repository-index digest.
`repository_symbol_graph`	read	Trace freshness-checked static calls without guessing symbol identity.
`repository_tests_for_changes`	read	Rank tests related to known changed paths without executing them.
`repository_verified_context`	read	Return a partial code graph scoped to one task with a receipt.
`resume_state`	read	Resume from the latest checkpoint.
`retrieve_compressed_span`	read	Retrieve one in-scope span and debit returned tokens from savings.
`scan_for_vulnerabilities`	read	Scan code for security vulnerabilities (SAST).
`search_compressed_spans`	read	Search in-scope spans and return bounded exact excerpts.
`security_report`	read	Session-wide security audit across all ingested fragments.
`security_scan`	read	Scan content for prompt injection attacks and security threats.
`shared_memory_forget`	destructive	Remove a shared memory entry by ID.
`shared_memory_list`	read	List shared memory entries, optionally filtered by agent or tag.
`shared_memory_search`	read	Search cross-agent shared memory by relevance. Returns entries from all agents.
`shared_memory_stats`	read	Get cross-agent shared memory statistics.
`shared_memory_write`	write	Write to cross-agent shared memory. Deduplicates near-identical entries via SimHash.
`smart_read`	read	Read a file at an automatic or caller-chosen resolution.
`start_workspace_listener`	write	Start a background workspace listener that continuously feeds repo changes into CogOps belief CI.
`steer_output`	read	Classify query effort and return output steering directives + max_tokens budget.
`sync_workspace_changes`	write	Synchronize workspace file changes into the belief and verification layers. Detects new/modified/deleted files, marks beliefs stale, recompiles.
`vault_hygiene_scan`	read	Scan vault beliefs against each other for knowledge decay.
`vault_query`	read	Query the CogOps Knowledge Vault for existing beliefs. Supports lookup by entity name or listing all.
`vault_search`	read	Full-text search across all belief artifacts in the vault. Uses keyword matching with entity-name boosting.
`vault_status`	read	Show the current state of the CogOps Knowledge Vault: total beliefs, verification status, confidence distribution, and routing statistics.
`vault_time_travel`	read	Query the vault
`vault_write_action`	write	Write a task output or report to the CogOps Knowledge Vault (actions/).
`vault_write_belief`	write	Write a belief artifact to the CogOps Knowledge Vault with machine-auditable frontmatter.
`verify_and_repair`	read	Verify LLM-generated code and suggest repairs for hallucinations.
`verify_beliefs`	write	Run a full verification pass on all beliefs. Checks staleness, contradictions, confidence divergence, low confidence scores.
`verify_provenance`	read	Verify that LLM-generated code is grounded in the provided context.
`verify_response`	read	Verify an AI-generated response for hallucination. Computes entity coverage gap, hedging curvature, and entropy consistency. Returns fused_risk [0.0=safe, 1.0=hallucinated], verdict (pass/warn/flag), and flagged claims. All local — zero LLM calls.
`work_acknowledge_recovery`	read	Accept responsibility for recovered work state so acting is allowed.
`work_claim`	read	Record explicit agent work plus a bounded advisory scope lease.
`work_compile_context`	read	Compile verified code context and record its Work Graph receipt.
`work_context_fault`	read	Fault exact omitted code from a context token or verified context object.
`work_handoff`	write	Create a graph-bound handoff receipt and complete continuation proof.
`work_modifications`	read	Modifications recorded between observations by the workspace watcher.
`work_record_context`	read	Attach a canonical ContextReceipt to its exact WorkScope.
`work_record_execution`	read	Atomically record route, observable execution and exact-head verification.
`work_record_memory`	read	Attach provenance-bearing memory without trusting raw model prose.
`work_resume`	read	Recover unfinished work and optionally seal a no-handoff proof.
`work_session_status`	read	Report the automatic takeover performed when this server started.
`work_state`	read	Inspect persisted shared work state without appending a polling event.

Trust audit

BLOCKgrade F · trust 41/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.

Layer	What it checks	Result
L0	Provenance & inventory	WARN
L1	Static analysis of the code	FAIL
L2	Instruction surface (what it tells the agent)	PASS
L3	Class-specific surface	WARN
L4	Behavioural (sandbox)	SKIPPED

What the source does

Filesystem: declared (7 observation(s))
Network: declared (8 observation(s))
Shell: declared (18 observation(s))
Dependencies: not all pinned
Secrets in source: found

Findings (25)

CRITICALHard-coded secrets · secret.private_key · CWE-798, CWE-321

entroly-engine/src/guardrails.rs:244

if lower.contains("-----begin private key-----")

CRITICALHard-coded secrets · secret.private_key · CWE-798, CWE-321

entroly-engine/src/guardrails.rs:245

|| lower.contains("-----begin rsa private key-----")

CRITICALHard-coded secrets · secret.private_key · CWE-798, CWE-321

entroly-engine/src/guardrails.rs:575

assert!(has_safety_signal("-----BEGIN PRIVATE KEY-----\nabc"));

CRITICALHard-coded secrets · secret.private_key · CWE-798, CWE-321

entroly-engine/src/guardrails.rs:576

assert!(has_safety_signal("-----BEGIN RSA PRIVATE KEY-----\nabc"));