@tsuga/cli@1.126.0
Tsuga CLI - manage resources from the command line
AI Security Review
scanned 2d ago · by lpm-firewall-aiThe bundle implements opt-in runtime telemetry for Tsuga’s own CLI-run endpoint, so this is not evidence of malware. Captured spans include request URL query strings, which may hold user queries or filters, and the generic consent prompt does not disclose that collection detail; warn as telemetry.
Static reason
One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review; source fingerprint signature matched known malicious package; routed for review
Trigger
After a successful CLI run, when telemetry is enabled by saved consent or TSUGA_TELEMETRY=1/true.
Impact
Potential disclosure of query or filter text to Tsuga’s API beyond what the generic usage-data prompt makes clear; no credential theft or unrelated collector was identified.
Mechanism
The CLI collects run metadata and spans, then starts a detached Node process to POST them to the configured Tsuga API’s /v1/cli-runs endpoint. HTTP spans record URL query strings.
AI rationale
The bundle implements opt-in runtime telemetry for Tsuga’s own CLI-run endpoint, so this is not evidence of malware. Captured spans include request URL query strings, which may hold user queries or filters, and the generic consent prompt does not disclose that collection detail; warn as telemetry.
Evidence
dist/cli.js
Network endpoints2
api.tsuga.com/v1/cli-runsConfigured Tsuga API base URL + /v1/cli-runs
Decision evidence
public snapshotAI called this Suspicious at 84.0% confidence as Dangerous Capability with medium false-positive risk.
Evidence for warning
- The CLI asks to send usage data on successful interactive runs unless the user has already set a telemetry preference; it saves the answer.
- The CLI sends run metadata and captured spans to its Tsuga API endpoint after consent. HTTP spans can include URL query strings, which may contain user queries or filters.
- The consent prompt says “Send usage data?” but does not explain that captured spans can include request URLs and query strings.
Evidence against
- The package has no install, preinstall, or postinstall hook; its executable is the CLI entry point.
- Telemetry is opt-in through a prompt or TSUGA_TELEMETRY setting, and the API describes the endpoint as recording Tsuga CLI runs.
- The code limits recorded command information to recognized CLI command names and hashes recognized AI-agent session IDs.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetwork
HighEntropyStrings
Source & flagged code
18 flaggeddist/cli.jsView file
273patternName = github_pat
severity = critical
line = 273
Critical
273patternName = github_pat
severity = critical
line = 273
Critical
273patternName = github_oauth
severity = critical
line = 273
Critical
273patternName = github_server_token
severity = critical
line = 273
Critical
Secret Pattern
GitHub server-to-server token in dist/cli.js
dist/cli.jsFindings
9 Critical9 High2 Medium4 Low
CriticalCritical Secretdist/cli.js
CriticalSecret Patterndist/cli.js
CriticalSecret Patterndist/cli.js
CriticalSecret Patterndist/cli.js
CriticalSecret Patterndist/cli.js
CriticalSecret Patterndist/cli.js
CriticalSecret Patterndist/cli.js
CriticalSecret Patterndist/cli.js
CriticalSecret Patterndist/cli.js
HighChild Processdist/cli.js
HighSame File Env Network Executiondist/cli.js
HighCommand Output Exfiltrationdist/cli.js
HighTrigger Reachable Command Output Exfiltration
Affected versions and remediation
This report applies to @tsuga/cli@1.126.0.
See all versions of @tsuga/cli for other recorded verdicts.
- Review the evidence and your use of @tsuga/cli@1.126.0 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.