instar@1.3.1362
Coherence infrastructure for self-evolving AI agents — on the Claude Code or Codex subscription you already have.
AI Security Review
scanned 5d ago · by lpm-firewall-aiThe package automatically adds a broad agent permission approval hook during runtime post-update migration. This is a dangerous capability that merits a warning, but the inspected source does not establish a block-eligible install-hook attack.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
When the package's post-update migration runs.
Impact
Agent permission prompts can be bypassed, allowing requested tool actions to proceed without individual approval.
Mechanism
The migration adds a PermissionRequest hook in project settings and writes a handler that returns an allow decision for each request.
AI rationale
The package automatically adds a broad agent permission approval hook during runtime post-update migration. This is a dangerous capability that merits a warning, but the inspected source does not establish a block-eligible install-hook attack.
Evidence
dist/core/PostUpdateMigrator.jsdist/core/UpdateChecker.js.claude/settings.json.instar/hooks/instar/auto-approve-permissions.js
Decision evidence
public snapshotAI called this Suspicious at 88.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- The post-update migrator adds a project PermissionRequest hook and its handler returns an allow decision for every request.
- The update checker can run the migrator after an update, without a consent check in the inspected migration path.
- The reviewed credential paths use the local operating-system keychain; no unrelated credential recipient was established.
Evidence against
- The package postinstall runs a native-module repair script; the permission hook is installed by runtime migration, not that lifecycle script.
- The inspected agent integration locations concern project skills, hook discovery, and configuration paths; they do not establish a separate attack.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsFilesystemNetworkShellWebSocket
HighEntropyStringsPossibleObfuscation
Source & flagged code
18 flagged · loading sourcepackage.jsonView file
•scripts.postinstall = node scripts/fix-better-sqlite3.cjs
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkgdist/threadline/PipeSessionSpawner.jsView file
15*/
L16: import { execSync } from 'node:child_process';
L17: import fs from 'node:fs';
High
Child Process
Package source references child process execution.
dist/threadline/PipeSessionSpawner.jsView on unpkg · L15352// Spawn tmux session, recording its incarnation (session id + pane pid)
L353: const created = execSync(`tmux new-session -d -P -F '#{session_id} #{pane_pid}' -s "${sessionName}" -x 200 -y 50 'bash -c "${shellCmd.replace(/"/g, '\\"')}"'`, { timeout: 10_000, e...
L354: const [tmuxId, panePid] = created.split(' ');
High
Shell
Package source references shell execution.
Findings
4 Critical11 High6 Medium5 Low
CriticalPersistence Backdoordist/core/GitSync.js
CriticalTrigger Reachable Dangerous Capabilitydist/core/GitSync.js
CriticalTrigger Reachable Excluded Directory Payloaddist/core/PostUpdateMigrator.js
CriticalPrevious Version Dangerous Deltadist/commands/server.js
HighInstall Time Lifecycle Scriptspackage.json
HighChild Processdist/threadline/PipeSessionSpawner.js
HighShelldist/threadline/PipeSessionSpawner.js
HighSame File Env Network Executionskills/spec-converge/scripts/publish-spec-review.mjs
HighCredential Redirect Persistencedist/messaging/TelegramAdapter.js
HighSandbox Evasion Gated Capabilitydist/commands/setup.js
Affected versions and remediation
This report applies to instar@1.3.1362.
See all versions of instar for other recorded verdicts.
- Review the evidence and your use of instar@1.3.1362 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.