Atlas / MCP servers / bekirdag / docdex

docdexBLOCK

Name: docdex
Author: bekirdag

mcp/bekirdag/docdex

MCP server by bekirdag · Audited 2026-10-12

⚑ Report / re-scan Source ↗

A document index with MCP, http and client support. https://docdex.org/

Verdict

BLOCK

Grade

Trust score

47 /100

Exposed tools

117 87r · 23w · 7d

Transport

streamable-http

License

MIT

Stars

Overview

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

[](https://docdex.org)

Turn your repository into fast, private context that humans and AI can trust.

Docdex is a local-first indexer and search daemon for documentation and source code. It sits between your raw files and your AI assistant, providing deterministic search, code intelligence, and persistent memory without ever uploading your code to a cloud vector store.

⚡ Why Docdex?

Most AI tools rely on "grep" (fast but dumb) or hosted RAG (slow and requires uploads). Docdex runs locally, understands code structure, and gives your AI agents a persistent memory.

🚀 Features

📚 Document Indexing: Rank and summarize repo documentation instantly.
🧠 AST & Impact Graph: Search by function intent and track downstream dependencies (supports Rust, Python, JS/TS, Go, Java, C++, and more).
💾 Repo Memory: Stores project facts, decisions, and notes locally.
👤 Agent Memory: Remembers user preferences (e.g., "Use concise bullet points") across different repositories.
**🗂️ Conversatio

Read from source at commit f42e69d4a90bOBSERVED · 2026-10-12

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 (npm)

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

Exposed tools (117)

87 read · 23 write · 7 destructive. Blast radius: 7 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
`constraints`	read	Optional constraints (performance, compatibility)
`depth`	read	Depth of the walkthrough (quick or deep)
`docdex_ai_terminal_capture`	read	Record a normalized AI-terminal session summary into personal preferences and mind-clone inputs.
`docdex_ai_terminal_detect`	read	Detect supported AI-terminal generated-skill targets without persisting integration records.
`docdex_ai_terminal_events`	read	List normalized AI-terminal capture events.
`docdex_ai_terminal_integrations`	read	List configured AI-terminal integrations for automatic capture, generated-skill sync, and activation.
`docdex_ai_terminal_integrations_bootstrap`	write	Create or update Docdex-owned generated-skill roots for supported AI terminals.
`docdex_ai_terminal_status`	write	Show AI-terminal capture and generated-skill sync status.
`docdex_ai_terminal_sync_skills`	write	Promote stable mind-clone playbooks into generated skills and install low-risk validated skills into enabled AI terminals.
`docdex_ast`	read	Read Tree-sitter AST nodes for a file, including per-file outcome (ok/skipped/failed).
`docdex_batch_search`	write	Run batched search queries with deterministic truncation and per-query responses.
`docdex_capabilities`	read	Return the Docdex optional-feature capability contract for Codali integration.
`docdex_clone_context`	read	Compile a bounded personal-preferences clone context pack for a query.
`docdex_clone_directive`	read	Compile an auditable operator-clone directive checklist for an agent task.
`docdex_clone_evaluate`	read	Score clone-pack fidelity heuristics for a query.
`docdex_clone_explain`	read	Explain the evidence trace for a bounded clone context pack.
`docdex_clone_replay_dataset`	read	Generate replay dataset cases from executable operator routines for clone-prediction evaluation.
`docdex_clone_replay_evaluate`	read	Score whether a clone context pack predicts expected operator next-step categories.
`docdex_clone_replay_suite`	write	Run replay cases through clone directives and report aggregate category, routine, step, and approval metrics.
`docdex_conversation_delete`	destructive	Delete a conversation session and its derived artifacts from a repo or explicit global conversation namespace.
`docdex_conversation_export`	read	Export one conversation session with any linked diary entries from a repo or explicit global conversation namespace.
`docdex_conversation_hook`	read	Enqueue or synchronously process a conversation-memory hook action for a repo or explicit global conversation namespace.
`docdex_conversation_import`	write	Import a normalized conversation session, plain-text transcript, or native export format into repo-scoped conversation memory or an explicit global conversation namespace.
`docdex_conversation_list`	read	List conversation sessions with summaries and metadata for a repo or explicit global conversation namespace.
`docdex_conversation_prune`	write	Preview or apply conversation retention cleanup for a repo or explicit global conversation namespace using current config or explicit overrides.
`docdex_conversation_read`	read	Read a conversation session with ordered messages and derived artifacts from a repo or explicit global conversation namespace.
`docdex_conversation_redact`	destructive	Redact one conversation session and clear its derived wake-up artifacts in a repo or explicit global conversation namespace.
`docdex_conversation_search`	read	Search conversation archive summaries and raw message snippets for a repo or explicit global conversation namespace.
`docdex_dag_export`	read	Export a DAG session for a given session_id (json/text/dot).
`docdex_diary_read`	read	Read diary entries for an optional agent in a repo or explicit global conversation namespace.
`docdex_diary_write`	write	Write one diary entry for an agent in a repo or explicit global conversation namespace.
`docdex_file`	read	Read a file from the current repo (delegates to docdex_open); vars: {path}.
`docdex_files`	read	List indexed documents (rel_path/doc_id/token_estimate) for the current repo.
`docdex_get_profile`	read	Fetch profile preferences for an agent (profile memory).
`docdex_impact_diagnostics`	write	List unresolved dynamic import diagnostics by file.
`docdex_impact_graph`	read	Fetch impact graph traversal for a repo-relative file.
`docdex_incident_triage`	read	Triage a production issue using Docdex search + impact data.
`docdex_index`	read	Rebuild the index (or ingest specific files) for the current repo root; use after large changes or stale results.
`docdex_kg_clear`	destructive	Delete all conversation knowledge graph data in a repo or explicit global conversation namespace.
`docdex_kg_delete_edge`	destructive	Delete one graph edge together with its fact projection and evidence in a repo or explicit global conversation namespace.
`docdex_kg_delete_episode`	destructive	Delete one provenance episode and its graph edges/facts in a repo or explicit global conversation namespace.
`docdex_kg_entity_links`	read	Fetch code-facing links recorded for one graph entity in a repo or explicit global conversation namespace.
`docdex_kg_episode`	read	Fetch one provenance episode together with its graph edges and evidence in a repo or explicit global conversation namespace.
`docdex_kg_neighborhood`	read	Inspect graph edges adjacent to one entity or topic in a repo or explicit global conversation namespace.
`docdex_kg_query`	read	Query temporal knowledge facts extracted from conversation memory in a repo or explicit global conversation namespace.
`docdex_kg_rebuild`	read	Rebuild graph-side link projections and reindex SQLite structures for conversation knowledge in a repo or explicit global conversation namespace.
`docdex_kg_search_edges`	read	Search graph-native edges extracted from conversation memory in a repo or explicit global conversation namespace.
`docdex_kg_search_episodes`	read	Search provenance episodes extracted from conversation memory in a repo or explicit global conversation namespace.
`docdex_kg_search_nodes`	read	Search canonical graph entities extracted from conversation memory in a repo or explicit global conversation namespace.
`docdex_kg_timeline`	read	Read the temporal timeline for one entity or decision topic in a repo or explicit global conversation namespace.
`docdex_llm_diagnostics`	read	Return local LLM service detection/default-selection diagnostics and status messages.
`docdex_local_completion`	read	Offload a small local task to a local model and return the draft output.
`docdex_memory_layers`	read	Describe the six Docdex memory layers, their scope, storage, and recommended agent usage for the current repo or conversation namespace.
`docdex_memory_recall`	read	Recall memory items by semantic similarity (requires DOCDEX_ENABLE_MEMORY=1).
`docdex_memory_route`	write	Recommend which Docdex memory lanes to consult first or write to for a specific query, including core versus retrievable memory.
`docdex_memory_save`	write	Store a memory item (requires DOCDEX_ENABLE_MEMORY=1).
`docdex_memory_store`	read	Store a memory item (requires DOCDEX_ENABLE_MEMORY=1).
`docdex_onboarding`	read	Guide a new contributor through the repo with Docdex tools.
`docdex_open`	read	Read a file from the repo (optional line window); rejects paths outside the repo.
`docdex_personal_preferences_categories`	read	List personal-preferences categories and context policy.
`docdex_personal_preferences_claim_forget`	read	Forget one claim, redact its usable content, and rebuild snapshots.
`docdex_personal_preferences_claim_override`	read	Override one claim with an explicit corrected value.
`docdex_personal_preferences_claim_read`	read	Read one extracted personal-preference claim.
`docdex_personal_preferences_claim_review`	read	Approve, reject, or re-queue one extracted claim.
`docdex_personal_preferences_claims`	read	List extracted personal-preference claims with truth/origin filters.
`docdex_personal_preferences_delete`	destructive	Delete one captured personal-preferences session and its derived records.
`docdex_personal_preferences_export`	read	Export one capture or the whole personal-preferences store to JSON.
`docdex_personal_preferences_feedback`	write	Add a feedback event that influences future mind-clone context.
`docdex_personal_preferences_generated_skill`	read	Read one generated skill by skill id, slug, or portable skill name.
`docdex_personal_preferences_generated_skill_disable`	write	Disable one generated skill while preserving registry provenance.
`docdex_personal_preferences_generated_skill_events`	read	List registry audit events for generated skills.
`docdex_personal_preferences_generated_skill_install`	write	Install one generated skill into enabled Docdex-owned terminal roots.
`docdex_personal_preferences_generated_skill_rollback`	read	Roll back one generated skill to the previous rendered version when available.
`docdex_personal_preferences_generated_skill_validate`	read	Validate one generated skill
`docdex_personal_preferences_generated_skills`	read	List generated skills derived from personal preferences and mind-clone operator routines.
`docdex_personal_preferences_generated_skills_autopilot`	write	Run one generated-skills autopilot pass through registry, validation, and install policy.
`docdex_personal_preferences_generated_skills_preview`	read	Render registry-backed generated-skill candidates without installing them.
`docdex_personal_preferences_generated_skills_render`	read	Render registry-backed generated-skill candidates without installing them.
`docdex_personal_preferences_generated_skills_sync`	write	Promote stable operator playbooks into versioned generated skills and optionally install them to terminal integrations.
`docdex_personal_preferences_list`	read	List captured personal-preferences sessions.
`docdex_personal_preferences_mind_map`	read	Compile a queryable operator mind map from approved claims and synthesized routines.
`docdex_personal_preferences_operator_event_record`	write	Record one manual operator event such as a git/test/deploy/approval/action step.
`docdex_personal_preferences_operator_events`	read	List first-class operator events captured from commands, approvals, and planning artifacts.
`docdex_personal_preferences_operator_events_scan_artifacts`	read	Scan repo planning/progress/SDS markdown artifacts into first-class operator events.
`docdex_personal_preferences_playbooks`	read	Compile stable operator routines into SKILL.md-compatible playbook artifacts.
`docdex_personal_preferences_process`	read	Manually process queued captures with a local mcoda agent.
`docdex_personal_preferences_prune`	write	Preview or apply personal-preferences retention pruning.
`docdex_personal_preferences_purge`	destructive	Purge all personal-preferences captures and derived records.
`docdex_personal_preferences_read`	read	Read one captured personal-preferences session.
`docdex_personal_preferences_redact`	read	Redact raw transcript content for one captured session.
`docdex_personal_preferences_retention_policies`	read	List personal-preferences retention policies by memory lane.
`docdex_personal_preferences_review`	read	Approve, reject, or re-queue one derived personal-preferences record.
`docdex_personal_preferences_reviews`	read	List derived personal-preferences records by review status.
`docdex_personal_preferences_routine_explain`	read	Explain one synthesized operator routine with source claims grouped by step.
`docdex_personal_preferences_routine_read`	read	Read one synthesized operator routine by id or routine key.
`docdex_personal_preferences_routines`	read	List synthesized operator routines compiled from repeated personal-preference claims.
`docdex_personal_preferences_routines_rebuild`	write	Rebuild synthesized operator routines from the current claim set.
`docdex_personal_preferences_scan`	read	Scan supported local AI-client transcript roots and queue new personal-preferences captures.
`docdex_personal_preferences_search`	read	Search derived personal-preferences records.
`docdex_personal_preferences_snapshot_read`	read	Read one temporal identity snapshot.
`docdex_personal_preferences_snapshots`	read	List temporal identity snapshots compiled from claims.
`docdex_personal_preferences_snapshots_rebuild`	write	Rebuild temporal identity snapshots from the current claim set.
`docdex_personal_preferences_status`	read	Show personal-preferences storage and queue status.
`docdex_refactor_plan`	read	Plan a refactor using Docdex symbols and impact graph.
`docdex_repo_inspect`	read	Inspect how Docdex resolves repo identity (normalized path, fingerprint, and any shared-state mapping).
`docdex_rerank`	read	Rerank candidate hits with transparent score updates and deterministic fallback behavior.
`docdex_save_preference`	write	Store a global profile preference for an agent (profile memory, async evolution).
`docdex_search`	read	Search repo docs/code and return hits with rel_path/path, summary, snippet, and doc_id.
`docdex_stats`	write	Inspect index metadata: doc count, state dir, size on disk, and last update time.
`docdex_symbols`	read	Read the symbol extraction result for a file, including per-file outcome (ok/skipped/failed).
`docdex_tree`	read	Render a folder tree with standard excludes (git, node_modules, build artifacts).
`docdex_wakeup`	read	Assemble bounded startup memory from conversation history in a repo or explicit global conversation namespace for the active agent/query.
`docdex_web_research`	write	Run local search plus web discovery/fetch and return the combined response (requires web enabled).
`log_snippet`	read	Optional log snippet for context
`symptom`	read	Symptom or error summary
`target`	read	What you want to refactor (module, file, API)
`topic`	read	Area or concept to focus on (e.g., auth, billing)

Trust audit

BLOCKgrade F · trust 47/100 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	PASS
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 (6 observation(s))
Network: declared (10 observation(s))
Shell: declared (6 observation(s))
Dependencies: not all pinned
Secrets in source: found

Findings (25)

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

src/cli/commands/self_check.rs:17

"-----BEGIN PRIVATE KEY-----",

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

src/daemon.rs:1535

const RSA_PKCS1_KEY: &str = r#"-----BEGIN PRIVATE KEY-----

HIGHNetwork egress · net.metadata_ip · CWE-200, CWE-319

src/orchestrator/web.rs:5455

assert!(!is_allowed_url("https://metadata.google.internal/", &[]));

Why it matters. cloud metadata endpoint: the classic SSRF credential grab

HIGHNetwork egress · net.metadata_ip · CWE-200, CWE-319

src/web/ddg.rs:2521

assert!(!is_url_allowed("https://metadata.google.internal/", &[]));

Why it matters. cloud metadata endpoint: the classic SSRF credential grab

HIGHNetwork egress · net.metadata_ip · CWE-200, CWE-319

src/web/policy.rs:434

"https://metadata.google.internal/",

Why it matters. cloud metadata endpoint: the classic SSRF credential grab

HIGHNetwork egress · net.metadata_ip · CWE-200, CWE-319

src/web/policy.rs:454

"http://169.254.169.254/latest/meta-data/",

Why it matters. cloud metadata endpoint: the classic SSRF credential grab

HIGHNetwork egress · net.tls_off · CWE-200, CWE-319

src/daemon.rs:472

"refusing to bind on non-loopback without TLS; provide --tls-cert/--tls-key or --insecure to allow plain HTTP",

Why it matters. certificate verification is disabled

Fix. leave verification on

HIGHNetwork egress · net.tls_off · CWE-200, CWE-319

src/daemon.rs:474

.with_hint("Provide `--tls-cert/--tls-key`, use `--certbot-domain/--certbot-live-dir`, or (unsafe) pass `--insecure` behind a trusted proxy.")

Why it matters. certificate verification is disabled

Fix. leave verification on

MEDIUMCode injection · code.dynamic_import · CWE-78, CWE-94, CWE-95

src/impact/parser.rs:3582

importlib.import_module(BASE + ".extra")

MEDIUMFilesystem / path · mcp.destructive_tools · CWE-22, CWE-59

docdex_conversation_delete, docdex_conversation_redact, docdex_kg_clear, docdex_kg_delete_edge, docdex_kg_delete_episode, docdex_personal_preferences_delete, docdex_personal_preferences_purge

Why it matters. 7 tool(s) can delete or overwrite

Fix. prefer a read-only mode or scoped tokens; the page states the blast radius

LOWInventory / provenance · inv.binary · CWE-1104

npm/test/fixtures/archive/fake-archive.bin

fake-archive.bin

Why it matters. a compiled or binary member cannot be reviewed from source

Fix. ship source, or explain the binary in the README

LOWInventory / provenance · inv.hidden_file · CWE-1104

.release-please-manifest.json

.release-please-manifest.json

Why it matters. hidden member outside the usual dotfiles

Fix. review its purpose

LOWCode injection · code.dynamic_import · CWE-78, CWE-94, CWE-95

tests/impact_graph_indexing.rs:51

importlib.import_module(BASE + ".extra")

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

npm/test/platform.test.js:20

const { DEFAULT_TARGETS } = require("../../scripts/generate_release_manifest.cjs");

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

npm/test/release_asset_validation.test.js:10

const { collectReleaseAssetIssues } = require("../../scripts/validate_release_assets.cjs");

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

npm/test/release_manifest_generator.test.js:14

} = require("../../scripts/generate_release_manifest.cjs");

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

src/llm/local_library.rs:4939

"../../tests/fixtures/local_llm_current_machine.json"

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

src/mcp/context/assembly.ts:2

import { getWorkspaceStore } from '../../core/workspace';