Loading npm security reports…
Loading npm security reports…
@strvmarv/total-recall@4.5.0: AI-agent control-surface risk | LPM Firewall
Soft block: AI-agent control surface
Warn by default; block when configured. Changes agent behavior and causes it to store user-provided information as memory without explicit consent.
AI security reviewscanned on Nov 12, 2026, 00:17 UTC · by lpm-firewall-ai The automatic postinstall writes agent configuration and Kiro steering instructions outside the package, changing agent behavior without a user command. The installed instructions direct ongoing capture of user information without asking permission, so this is an AI-agent control hijack.
Static reason
High-risk behavior combination matched malicious policy.; source matched previously finalized malicious package; routed for review
Trigger
npm postinstall when a supported agent is detected in the user's environment.
Impact
Changes agent behavior and causes it to store user-provided information as memory without explicit consent.
Mechanism
Registers total-recall as an MCP server and, for Kiro, copies an always-included steering file directing the agent to capture user corrections, preferences, and decisions without asking.
Policy narrative
Installation automatically configures detected AI-agent hosts to run total-recall. For Kiro, it also installs always-included steering instructions that tell the agent to capture user corrections, preferences, and decisions without asking permission and to search stored memories on each task. This unrequested change to an external agent's behavior is an AI-agent control hijack.
Evidence
package.json scripts/postinstall.js scripts/install-kiro-hooks.js scripts/lib/merge-json-config.js
.kiro/steering/total-recall.md
~/.kiro/settings/mcp.json
~/.kiro/steering/total-recall.md
~/.vscode/cline_mcp_settings.json
~/.codeium/windsurf/mcp_config.json
~/.codeium/windsurf/rules/total-recall.md
Decision evidencepublic snapshot AI identified this as a dangerous AI-agent capability at 97.0% confidence with low false-positive risk.
Evidence for policy risk
The package runs scripts/postinstall.js automatically after installation. Postinstall invokes agent setup routines, including Kiro, Cline, Windsurf, Gemini, Zed, and VS Code setup. When Kiro is detected, the installer registers the package as an MCP server and copies its steering instructions into Kiro's user-level steering directory. The installed steering instructs the agent to store user corrections, preferences, and decisions without asking permission, and to search memory on each question or task. Evidence against
The setup skips hosts it does not detect and avoids replacing an existing MCP entry. The package describes itself as a local memory tool, and the downloaded engine comes from its GitHub Releases host. Behavioral surface
Source Child process · Crypto · Dynamic require · Environment variables · Filesystem · Network · Shell
Supply chain High entropy strings · URL strings
Manifest No manifest risk signals triggered.
Scanned 78 files · 514 KB of source
2 external domains referenced
Findings1 critical · 15 high · 7 medium · 5 low Source & flagged code20 flagged High Install time lifecycle scripts
Package defines install-time lifecycle scripts.
package.json View on unpkg Medium Ambiguous install lifecycle script
Install-time lifecycle script is not statically allowlisted and needs review.
package.json View on unpkg Shared scanner record · 2 findings · 1 line scripts.postinstall = node scripts/postinstall.js
Medium Dynamic require
Package source references dynamic require/import behavior.
copilot/hook.js View on unpkg · L133 Code excerpt · 3 lines 133 readFile: (p) => fs.readFileSync (p, 'utf8' ),
134 loadFloor: () => import (pathToFileURL (path.join (rootDir, 'pi' , 'floor.js' )).href),
135 });
Critical AI agent control hijack
Source creates an unconsented AI-agent control surface through install-time mutation or a default unauthenticated remote skill channel.
scripts/install-cline.js View on unpkg · L30 Code excerpt · 16 lines Install-time code directly mutates a foreign AI-agent control surface:
30 const configPath = join (chome, 'cline_mcp_settings.json' );
31 mergeJsonConfig (configPath, 'mcpServers' , {
32 command: 'npx' ,
Write operation from scripts/fetch-binary.js:
198 export function writeVerifiedMarker (rid, version, sha256) {
199 try { fs.writeFileSync (verifiedMarkerPath (rid), JSON.stringify ({ version, sha256 })); } { }
High Trigger reachable external AI agent control surface mutation
Manifest-trigger-reachable source writes behavior-bearing configuration into a user or project AI-agent control surface.
scripts/install-kiro-hooks.js View on unpkg Scanner record · 17 lines Manifest-trigger-reachable source links an external AI-agent control path to a behavior-bearing write operation.
binaries/linux-arm64/runtimes/vec0.so View file Medium Ships native binary
Package ships native binary artifacts.
binaries/linux-arm64/runtimes/vec0.so View on unpkg Scanner record · 4 lines path = binaries/linux-arm64/runtimes/vec0.so
kind = native_binary
sizeBytes = 156520
magicHex = [redacted]
Medium Ships build helper
Package ships non-JavaScript build or shell helper files.
hermes-plugin/provider.py View on unpkg Scanner record · 4 lines path = hermes-plugin/provider.py
kind = build_helper
sizeBytes = 19597
magicHex = [redacted]
High Semantic analysis limited
A bounded semantic-analysis stage reached its safety limit; remaining detectors completed, but this package requires AI review.
codex/manifest.test.js View on unpkg Scanner record · 1 line stage = ast_semantic_analysis; reason = ast_trace_path_limit_exceeded; limitedFiles = 2
High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
scripts/fetch-binary.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = scripts/fetch-binary.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
codex/hook.test.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = codex/hook.test.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
cursor/hook.test.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = cursor/hook.test.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
skills/develop/scripts/ghpin.test.js View file High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
skills/develop/scripts/ghpin.test.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = skills/develop/scripts/ghpin.test.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
copilot/hook.test.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = copilot/hook.test.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
pi/floor.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = pi/floor.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
scripts/install-hermes-plugin.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = scripts/install-hermes-plugin.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
skills/develop/scripts/ghpin.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = skills/develop/scripts/ghpin.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
pi/extension.test.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = pi/extension.test.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
pi/floor.test.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = pi/floor.test.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
skills/develop/scripts/args.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = skills/develop/scripts/args.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High Known malware source similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
skills/develop/scripts/args.test.js View on unpkg Scanner record · 6 lines matchType = normalized_sha256
matchedPackage = @strvmarv/total-recall@4.4.0
matchedPath = skills/develop/scripts/args.test.js
matchedIdentity = npm:QHN0cnZtYXJ2L3RvdGFsLXJlY2FsbA:4.4.0
similarity = 1.000
summary = normalized source hash matched finalized malicious source
AI assessment confidence 97%
Certainty of the verdict, not severity.
Verdict details False-positive risk, Review stage, Firewall policy This finding concerns changes to an AI agent's instructions or configuration without explicit user action. It does not by itself establish malware intent. The CLI warns by default and blocks when configured for this policy.
Package metadata npm Author strvmarv
Version 4.5.0
Published Nov 12, 2026
License MIT
Dependencies 0
Integrity verified
Package size 157 MB
Files 163
Keywords mcp memory agent-memory knowledge-base rag claude-code pi-package codex
Runtime surface package.json Entrypoints main → bin/start.js bin → total-recall
Runtime node >=20.0.0
Artifact 342 MB unpacked · 163 files 2 signatures · provenance
Install lifecycle postinstall prepare prepublishOnly
On this report AI security review Decision evidence Findings28 Flagged code20 Remediation High Known malware source similarity Code↓
Package metadata 8 findings
Medium Network
Medium Environment variables
Medium Structural risk force deep review
Low Non install lifecycle scripts
catch
/* best-effort */
328 fs.mkdirSync (destDir, { recursive: true });
476 const lockPath = provisionLockPath (rid);
477 try { fs.mkdirSync (path.dirname (lockPath), { recursive: true }); } catch {}
478 if (fs.existsSync (lockPath) && provisionerAlive (lockPath)) return { status: 'in-progress' , rid };
import { cpsync, mkdirsync, existssync } from 'node:fs';
import { homedir } from 'node:os';
import { join } from 'node:path';