registry / @punks/cli / 6.1.1
@punks/cli@6.1.1
Harness Intelligence scaffolding CLI
AI Security Review
scanned 1mo ago · by lpm-firewall-aiThe package is not malicious by the inspected evidence, but its automatic project-level agent configuration is a concrete extension-lifecycle capability. It should be warned rather than blocked.
Static reason
One or more suspicious static signals were detected.
Trigger
A user runs a CLI registry workflow that reaches semantic reads.
Impact
It can alter agent configuration in the selected project; no broader control-surface mutation or data export was established.
Mechanism
The workflow writes a Typegraph configuration block to the project-local Codex configuration file.
AI rationale
The package is not malicious by the inspected evidence, but its automatic project-level agent configuration is a concrete extension-lifecycle capability. It should be warned rather than blocked.
Evidence
dist/index.js.codex/config.toml
Decision evidence
public snapshotAI called this Suspicious at 86.0% confidence as Dangerous Capability with low false-positive risk.
Evidence for warning
- The package exposes command-line binaries and has no npm install lifecycle hook in its inspected manifest.
- The CLI can write a managed Typegraph block to the project's .codex/config.toml.
- The registry application workflow invokes the semantic-reads step, which can perform that configuration update.
Evidence against
- No credential collection, exfiltration sink, hidden remote payload execution, or install-time hook was identified in the inspected entrypoint.
Behavioral surface
ChildProcessCryptoDynamicRequireEnvironmentVarsEvalFilesystemNetworkShell
HighEntropyStringsMinifiedPossibleObfuscationUrlStrings
NoLicenseWildcardDependency
Source & flagged code
9 flaggeddist/index.jsView file
155`)+1;while(nl!==0){this.onNewLine(this.offset+nl);nl=this.source.indexOf(`
L156: `,nl)+1}}return{type,offset:this.offset,indent:this.indent,source:this.source}}startBlockValue(parent){switch(this.type){case"alias":case"scalar":case"single-quoted-scalar":case"do...
L157: `+ind(d+1));return`[
High
349`);const preCommit=lines2.findIndex((line)=>line.trim()==="pre-commit:");if(preCommit===-1){return null}const preIndent=indent(lines2[preCommit]);const preCommitEnd=lines2.findInde...
L350: `).some((line)=>{const command=line.trim();return command.length>0&&!command.startsWith("#")&&(/^(?:(?:command|exec)\s+)?lefthook\s+run\s+["']?pre-commit["']?(?:\s|$)/u.test(comman...
L351:
High
151`)return yield*this.pushCount(1);else if(ch==="\r"&&this.charAt(1)===`
L152: `)return yield*this.pushCount(2);else return 0}*pushSpaces(allowTabs){let i=this.pos-1;let ch;do{ch=this.buffer[++i]}while(ch===" "||allowTabs&&ch==="\t");const n=i-this.pos;if(n>0...
L153: `)+1;while(nl!==0){this.onNewLine(this.offset+nl);nl=this.source.indexOf(`
...
L155: `)+1;while(nl!==0){this.onNewLine(this.offset+nl);nl=this.source.indexOf(`
L156: `,nl)+1}}return{type,offset:this.offset,indent:this.indent,source:this.source}}startBlockValue(parent){switch(this.type){case"alias":case"scalar":case"single-quoted-scalar":case"do...
L157: `+ind(d+1));return`[
Findings
7 High6 Medium6 Low
HighChild Processdist/index.js
HighShelldist/index.js
HighSame File Env Network Executiondist/index.js
HighCredential Redirect Persistencedist/index.js
HighCommand Output Exfiltrationdist/index.js
HighObfuscated Payload Loaderdist/index.js
HighTrigger Reachable Command Output Exfiltrationdist/index.js
MediumDynamic Requiredist/index.js
MediumUnsafe Vm Contextdist/index.js
MediumNetwork
MediumEnvironment Vars
MediumStructural Risk Force Deep Review
MediumWildcard Dependency
LowScripts Present
Affected versions and remediation
This report applies to @punks/cli@6.1.1.
See all versions of @punks/cli for other recorded verdicts.
- Review the evidence and your use of @punks/cli@6.1.1 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.