registry / pi-harness-runtime / 0.3.0

pi-harness-runtime@0.3.0

Codex-style /usage status for pi: local token tracking + provider mirror

AI Security Review

scanned 1mo ago · by lpm-firewall-ai

No install-time or covert exfiltration attack surface is established. Exported harness utilities retain dangerous explicit-use primitives: shell-backed git worktree commands and browser-page string evaluation.

Static reason

No blocking static signals were detected.

Trigger

A consumer explicitly instantiates `WorktreeManager` with attacker-controlled options or runs E2E assertions through `PlaywrightE2ERunner`.

Impact

Potential command execution in the configured repository context or script execution in the navigated browser page.

Mechanism

shell command construction and Playwright page-context eval

Rationale

Source inspection found no lifecycle persistence, credential collection, network exfiltration, or hidden payload chain. Explicit-use command and eval primitives create real unresolved security risk, so a warning is warranted rather than a publication block.

Evidence

package.jsonindex.tscli.tspackages/worktree/worktree.tsharness/e2e/playwright-runner.ts~/.pi/usage-status/usage.jsonl~/.pi/usage-status/mirror.json~/.pi/harness

Decision evidence

public snapshot

AI called this Suspicious at 90.0% confidence as Dangerous Capability with low false-positive risk.

pi-harness-runtime@0.3.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code