registry / @plannotator/opencode / 0.27.21
@plannotator/opencode@0.27.21
Plannotator plugin for OpenCode - interactive plan review with visual annotation
AI Security Review
scanned 28d ago · by lpm-firewall-aiThe postinstall hook changes OpenCode configuration without an explicit command, so this warrants a warning under the first-party agent extension setup policy. The inspected behavior is package-owned setup aligned with the plugin, with no confirmed attack surface beyond that lifecycle mutation.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
Installing the package with npm, which runs postinstall.
Impact
The package adds Plannotator commands and skill guidance to the user's OpenCode configuration.
Mechanism
The hook creates OpenCode commands and skill directories and copies the package’s command and skill files there.
AI rationale
The postinstall hook changes OpenCode configuration without an explicit command, so this warrants a warning under the first-party agent extension setup policy. The inspected behavior is package-owned setup aligned with the plugin, with no confirmed attack surface beyond that lifecycle mutation.
Evidence
package.jsoncommands/plannotator-review.mdskills/plannotator/SKILL.md
Decision evidence
public snapshotAI called this Suspicious at 86.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- The postinstall script copies package commands and a skill into the user’s OpenCode configuration directories.
- The installed review command asks the agent to run Plannotator and relay its output; this matches the package’s review workflow.
- The server dynamically imports its bundled embedded runtime as part of package operation.
Evidence against
- The install-time writes are limited to OpenCode commands and a Plannotator skill, consistent with this OpenCode plugin’s purpose.
- Inspected runtime excerpts show bundled-runtime loading, not a remote payload or credential transfer.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsEvalFilesystemNetworkShellWebSocket
HighEntropyStringsUrlStrings
Source & flagged code
5 flaggedpackage.jsonView file
•scripts.postinstall = mkdir -p ${XDG_CONFIG_HOME:-$HOME/.config}/opencode/commands && cp ./commands/*.md ${XDG_CONFIG_HOME:-$HOME/.config}/opencode/commands/ 2>/dev/null || true; mkdir -p ${XDG_CONFIG_H...
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg•scripts.postinstall = mkdir -p ${XDG_CONFIG_HOME:-$HOME/.config}/opencode/commands && cp ./commands/*.md ${XDG_CONFIG_HOME:-$HOME/.config}/opencode/commands/ 2>/dev/null || true; mkdir -p ${XDG_CONFIG_H...
Medium
Ambiguous Install Lifecycle Script
Install-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgdist/server.jsView file
•matchType = previous_version_dangerous_delta
matchedPackage = @plannotator/opencode@0.27.14
matchedIdentity = npm:QHBsYW5ub3RhdG9yL29wZW5jb2Rl:0.27.14
similarity = 0.500
summary = stored previous version shares package body but lacks this dangerous source file
Findings
2 High6 Medium6 Low
HighInstall Time Lifecycle Scriptspackage.json
HighPrevious Version Dangerous Deltadist/server.js
MediumAmbiguous Install Lifecycle Scriptpackage.json
MediumDynamic Requiredist/server.js
MediumNetwork
MediumEnvironment Vars
MediumOversized Source Filedist/embedded.js
MediumStructural Risk Force Deep Review
LowNon Install Lifecycle Scripts
LowScripts Present
LowEval
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
Affected versions and remediation
This report applies to @plannotator/opencode@0.27.21.
See all versions of @plannotator/opencode for other recorded verdicts.
- Review the evidence and your use of @plannotator/opencode@0.27.21 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.