Portable MCP server giving AI agents persistent, verified, cross-session memory
Inferred from the transports this listing declares (stdio). A client not listed here hasn’t been ruled out — it just isn’t something Forge can confirm.
Verification confirms publisher identity (repo ownership), not code safety. The security scan covers known CVEs and suspicious install scripts.
Read out of the source npm actually ships, at scan time. The package was never executed. Tools registered dynamically at runtime, or hidden inside bundled or minified code, can be missed — so this is a floor on the tool surface, not a complete census of it.
load_contextLoad relevant knowledge entries based on a message and project context. Uses smart loading with relevance scoring, trigger matching, and context budget management.Load relevant knowledge entries based on a message and project context. Uses smart loading with relevance scoring, trigger matching, and context budget management.
No input schema was published for this tool.
get_budgetGet current context budget usage and availability. Shows total tokens, allocation by slice, and remaining capacity.Get current context budget usage and availability. Shows total tokens, allocation by slice, and remaining capacity.
No input schema was published for this tool.
correctRecord a correction — the highest-priority knowledge type. Creates a new correction entry that supersedes an existing knowledge entry. The original is deactivated. Corrections are ALWAYS loaded at session start, never archived.Record a correction — the highest-priority knowledge type. Creates a new correction entry that supersedes an existing knowledge entry. The original is deactivated. Corrections are ALWAYS loaded at session start, never archived.
No input schema was published for this tool.
delegateCreate a delegation package for handing off a task to another agent with full context. Includes relevant decisions, errors, patterns, and constraints.Create a delegation package for handing off a task to another agent with full context. Includes relevant decisions, errors, patterns, and constraints.
No input schema was published for this tool.
add_dependencyAdd a dependency relationship between two components in a project. Used to track impact of changes.Add a dependency relationship between two components in a project. Used to track impact of changes.
No input schema was published for this tool.
query_dependenciesQuery the dependency graph for a component. Returns all components that depend on it or that it depends on, showing change impact.Query the dependency graph for a component. Returns all components that depend on it or that it depends on, showing change impact.
No input schema was published for this tool.
remove_dependencyprivilegedRemove a dependency relationship by its ID.Remove a dependency relationship by its ID.
No input schema was published for this tool.
rememberStore a new piece of knowledge in the Total Recall memory system. Use this to persist facts, decisions, errors, corrections, learnings, patterns, preferences, rules, risks, or any other knowledge that should survive across sessions.Store a new piece of knowledge in the Total Recall memory system. Use this to persist facts, decisions, errors, corrections, learnings, patterns, preferences, rules, risks, or any other knowledge that should survive across sessions.
No input schema was published for this tool.
recallSearch the knowledge base using a natural language query. Returns matching entries ranked by relevance. Use this to retrieve facts, decisions, errors, or any stored knowledge.Search the knowledge base using a natural language query. Returns matching entries ranked by relevance. Use this to retrieve facts, decisions, errors, or any stored knowledge.
No input schema was published for this tool.
recall_by_categoryRetrieve all active knowledge entries in a specific category. Useful for loading all decisions, all corrections, all errors, etc.Retrieve all active knowledge entries in a specific category. Useful for loading all decisions, all corrections, all errors, etc.
No input schema was published for this tool.
recall_by_idRetrieve a single knowledge entry by its UUID. Use when you have a specific entry ID.Retrieve a single knowledge entry by its UUID. Use when you have a specific entry ID.
No input schema was published for this tool.
forgetArchive a knowledge entry (soft-delete). The entry is preserved but no longer auto-loaded. Use when knowledge is outdated or superseded.Archive a knowledge entry (soft-delete). The entry is preserved but no longer auto-loaded. Use when knowledge is outdated or superseded.
No input schema was published for this tool.
recall_semanticSearch knowledge using semantic TF-IDF ranking. Better than keyword search for natural language questions like 'What did we decide about hosting?' or 'What errors happened with PowerShell?'Search knowledge using semantic TF-IDF ranking. Better than keyword search for natural language questions like 'What did we decide about hosting?' or 'What errors happened with PowerShell?'
No input schema was published for this tool.
get_suggestionsView pending knowledge suggestions from the behavioral learning engine. These are patterns Total Recall detected from your usage. Approve or dismiss them.View pending knowledge suggestions from the behavioral learning engine. These are patterns Total Recall detected from your usage. Approve or dismiss them.
No input schema was published for this tool.
approve_suggestionApprove a pending behavioral suggestion, promoting it to permanent knowledge.Approve a pending behavioral suggestion, promoting it to permanent knowledge.
No input schema was published for this tool.
dismiss_suggestionDismiss a pending behavioral suggestion. It won't be shown again.Dismiss a pending behavioral suggestion. It won't be shown again.
No input schema was published for this tool.
session_startStart a new Total Recall session. Call this at the beginning of every conversation. Returns session_id and pre-loaded corrections.Start a new Total Recall session. Call this at the beginning of every conversation. Returns session_id and pre-loaded corrections.
No input schema was published for this tool.
session_state_readRead the current session state. Call this before generating any response to refresh your memory of what happened earlier in this conversation.Read the current session state. Call this before generating any response to refresh your memory of what happened earlier in this conversation.
No input schema was published for this tool.
session_state_writeprivilegedUpdate the session state. Call this after every response that involves tool calls, file operations, decisions, or errors.Update the session state. Call this after every response that involves tool calls, file operations, decisions, or errors.
No input schema was published for this tool.
session_closeClose the current session. Call this at the end of every conversation. Triggers final state save, audit, and quality scoring.Close the current session. Call this at the end of every conversation. Triggers final state save, audit, and quality scoring.
No input schema was published for this tool.
health_checkRun a comprehensive health check on Total Recall. Returns database integrity status, table counts, session directory status, and machine info.Run a comprehensive health check on Total Recall. Returns database integrity status, table counts, session directory status, and machine info.
No input schema was published for this tool.
export_knowledgeExport all active knowledge and sessions to a JSON file for git backup or transfer.Export all active knowledge and sessions to a JSON file for git backup or transfer.
No input schema was published for this tool.
doctorRun diagnostic checks and repair common issues. Checks database integrity, orphaned session files, checksum mismatches, and WAL size.Run diagnostic checks and repair common issues. Checks database integrity, orphaned session files, checksum mismatches, and WAL size.
No input schema was published for this tool.
syncSync knowledge between local SQLite and Supabase cloud. Push local changes, pull remote changes, or both. Returns offline status if Supabase is not configured.Sync knowledge between local SQLite and Supabase cloud. Push local changes, pull remote changes, or both. Returns offline status if Supabase is not configured.
No input schema was published for this tool.
task_createCreate a multi-session task for tracking work across conversations.Create a multi-session task for tracking work across conversations.
No input schema was published for this tool.
task_updateUpdate a multi-session task's status, add timeline entries, or update blockers/deliverables/next_steps.Update a multi-session task's status, add timeline entries, or update blockers/deliverables/next_steps.
No input schema was published for this tool.
task_listList multi-session tasks, optionally filtered by project or status.List multi-session tasks, optionally filtered by project or status.
No input schema was published for this tool.
verify_fileVerify that a file exists, has content, and is optionally tracked in git. Returns VERIFIED, FAILED, or STALE.Verify that a file exists, has content, and is optionally tracked in git. Returns VERIFIED, FAILED, or STALE.
No input schema was published for this tool.
verify_pushVerify that local HEAD matches remote HEAD for a repository branch.Verify that local HEAD matches remote HEAD for a repository branch.
No input schema was published for this tool.
verify_claimVerify a claim against stored decisions and facts in the knowledge base. Returns VERIFIED, CONFLICT, or UNKNOWN.Verify a claim against stored decisions and facts in the knowledge base. Returns VERIFIED, CONFLICT, or UNKNOWN.
No input schema was published for this tool.
30 of 30 tools published a description.
Tool names and descriptions are written by the publisher and shown verbatim as inert text. They are the strings an MCP client passes to a model, so Forge scans them for prompt-injection patterns — any finding appears with the security scan above. “Privileged” is a keyword match on the tool name, not an audit of what the tool does: a benign-sounding name can still do anything.
Portable MCP server giving AI agents persistent, verified, cross-session memory
Linked names open Forge’s index of every entry observed exposing that tool. Browse all indexed tools.
The crawl stopped at the 60-package limit. The rest of the tree was never resolved.
36 more resolved packages are not drawn here (display cap: 24). Every dependency carrying an advisory is drawn regardless of the cap. Full inventory (CycloneDX SBOM)
104 declared dependencies never landed in the tree. They are missing from Forge's resolution, not from the package.
+92 more not listed. The counts by reason above cover all of them.
Not followed: peerDependencies, optionalDependencies. This tree covers runtime dependencies only, so anything those pull in was never resolved.