📦

agent-transit

Name: agent-transit
Author: israelvicars

Révision du contenu r1 Risque élevé 📁 Accès au système de fichiers⚡ Contient des scripts⚙️ Commandes externes🔑 Variables d’environnement🌐 Accès réseau

Visualize Coding Agent Sessions as Transit Maps

Coding session logs make tool sequences and delays difficult to review. Agent Transit turns one local session into an interactive map with timing and failure markers.

Prend en charge: Claude Codex Code(CC)

⚠️ 38 Médiocre

Maintain this Skill? Claim it ⚠️ 38 Médiocre

Arrêtez-vous et demandez confirmation avant l’installation.

Examinez le plan et obtenez le consentement explicite de l’utilisateur avant de modifier des fichiers.

Installer avec mon Agent

Copiez cette demande dans votre Agent. Elle contient la page Skill de référence et le manifeste.

Demande d’agent

Review the Skillstore skill "agent-transit" from https://skillstore.io/skills/israelvicars-agent-transit.md and its manifest at https://skillstore.io/api/skills/israelvicars-agent-transit/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.

Votre Agent doit toujours présenter son plan et demander toute confirmation exigée par la politique de sécurité.

Ressources lisibles par les agents

Utilisez ces liens lorsqu’un Agent IA, un crawler ou un script a besoin d’un contexte propre au lieu de lire toute la page.

Détail Markdown GET /skills/israelvicars-agent-transit.md manifest signé GET /api/skills/israelvicars-agent-transit/manifest lockfile signé GET /api/skills/israelvicars-agent-transit/lockfile

Tester

Utilisation de « agent-transit ». Illustrative request: map a saved Claude Code session containing a denied shell call.

Résultat attendu:

A local interactive HTML map grouped by tool family.
A service alert on the denied shell call.
A summary identifying the failed call without its command payload.

Utilisation de « agent-transit ». Illustrative request: animate a synthetic Codex session and verify the result.

Résultat attendu:

Local HTML, GIF, MP4, and contact-sheet artifacts.
Verification results for repeated frame hashes, departure timing, and blank-frame detection.

Audit de sécurité

Risque élevé

v1 • 22/10/2026 Ouvrir le rapport versionné

Most static alerts are false positives from JavaScript syntax, regular expressions, synthetic fixtures, or expected local tooling. Confirmed issues include optional contract HTML injection, plaintext credential transport, and strict-sharing privacy gaps. No evidence found of malicious intent; 278 unpresented static matches still require manual review before automatic publication. Static review was capped at 400/678 representative findings; omitted static matches are unconfirmed, so automatic publishing stays disabled until manual review.

Fichiers analysés

4,644

Lignes analysées

Éléments à examiner

Faux positifs ignorés

Préoccupations de sécurité confirmées (2)

Élevé

src/request-transit/adapters/normalize.mjs:406-408 src/request-transit/adapters/normalize.mjs:436-446 core/scrub.mjs:82-111 core/build.mjs:117

Internal trace passthrough bypasses strict sharing restrictions

Internal-format validation preserves extra fields and metadata. Strict scrubbing leaves those fields and nested event objects intact before embedding the entire trace in HTML.

A synthetic strict-share trace retained arbitrary span payloads, private metadata, and nested event messages. The builder serializes the complete retained object.

Moyen

src/agent-transit/adapters/normalize-agent.mjs:37-47 src/agent-transit/adapters/normalize-agent.mjs:195-197 SKILL.md:12

Strict sharing retains search argument fragments

Default tool labels include the first word of search queries, patterns, or URLs. Confidential query terms can survive strict scrubbing and enter shared HTML.

A synthetic WebSearch query retained its confidential first word after normalization and strict-share scrubbing. This occurs without --show-args.

Éléments d’évaluation des capacités (4)

Il s’agit de capacités locales réelles qui peuvent être attendues pour cette compétence ; elles nécessitent donc un examen, mais ne sont pas comptabilisées comme comportement malveillant confirmé.

Moyen

core/transit.template.html:994

innerHTML assignment (XSS risk)

panel.innerHTML =

With --contracts, card() inserts side.confidence without escaping into panel.innerHTML. A malicious index can inject active markup into the generated page.

Moyen

src/request-transit/adapters/fetch-posthog-api.mjs:28

Fetch API call

const r = await fetch(`${host}/api/projects/@current/`, { headers: { Authorization: `Bearer ${key}`

POSTHOG_HOST accepts HTTP endpoints, and this fetch sends the bearer key without enforcing HTTPS. A network observer could capture credentials.

Moyen

src/request-transit/adapters/fetch-posthog-api.mjs:36

Fetch API call

const r = await fetch(url, { method: 'POST', headers: { Authorization: `Bearer ${key}`, 'Content-Typ

POSTHOG_HOST accepts HTTP endpoints, and this fetch sends the bearer key without enforcing HTTPS. A network observer could capture credentials.

Moyen

src/request-transit/adapters/fetch-tempo.mjs:35

Fetch API call

try { return await fetch(fetchBase + path, { headers }); }

TEMPO_URL accepts HTTP, and get() forwards bearer or Basic authentication without enforcing HTTPS. Remote plaintext endpoints can expose credentials to network observers.

Facteurs de risque

📁 Accès au système de fichiers (50)

core/adapters.mjs:1 core/adapters.mjs:2 core/adapters.mjs:3 core/build.mjs:11 core/build.mjs:60 core/build.mjs:115 core/build.mjs:118 core/build.mjs:119 core/build.mjs:126 core/build.mjs:14 core/build.mjs:17 core/export.mjs:29 core/fsguard.mjs:5 core/fsguard.mjs:43 core/fsguard.mjs:48 core/scrub.mjs:25 core/scrub.mjs:31 core/scrub.mjs:33 core/scrub.mjs:98 core/setup.mjs:24 core/setup.mjs:36 core/setup.mjs:126 core/setup.mjs:160 core/setup.mjs:190 core/setup.mjs:58 core/setup.mjs:81 core/setup.mjs:101 core/still.mjs:9 core/tools.mjs:8 core/tools.mjs:15 core/tools.mjs:20 core/transit.template.html:914 core/transit.template.html:1095 core/verdict.mjs:96 core/verify.mjs:8 core/verify.mjs:35 core/verify.mjs:63 core/verify.mjs:65 src/agent-transit/adapters/normalize-agent.mjs:4 src/agent-transit/adapters/normalize-agent.mjs:209 src/agent-transit/adapters/normalize-agent.mjs:6 src/agent-transit/adapters/normalize-agent.mjs:107 src/agent-transit/samples/make-agent-samples.mjs:1 src/agent-transit/samples/make-agent-samples.mjs:30 src/agent-transit/samples/README.md:3 src/agent-transit/samples/README.md:3 src/agent-transit/scripts/export-cursor.mjs:18 src/agent-transit/scripts/export-cursor.mjs:52 src/agent-transit/scripts/export-cursor.mjs:113 src/agent-transit/scripts/export-cursor.mjs:37

⚡ Contient des scripts (15)

core/build.mjs:131 core/render.mjs:24 core/setup.mjs:92 core/still.mjs:3 core/tools.mjs:26 core/transit.template.html:804 core/transit.template.html:986 core/transit.template.html:987 core/transit.template.html:994 core/verify.mjs:17 src/agent-transit/scripts/export-cursor.mjs:38 src/request-transit/adapters/normalize.mjs:289 src/request-transit/adapters/normalize.mjs:293 src/request-transit/contracts/extractors/openapi.mjs:1 src/request-transit/contracts/extractors/protobuf.mjs:60

⚙️ Commandes externes (50)

core/build.mjs:53 core/build.mjs:61 core/build.mjs:62 core/build.mjs:66 core/build.mjs:67 core/build.mjs:73 core/build.mjs:80 core/build.mjs:95 core/build.mjs:96 core/build.mjs:114 core/build.mjs:117 core/build.mjs:118 core/build.mjs:124 core/build.mjs:137 core/build.mjs:138 core/cli.mjs:17 core/cli.mjs:32 core/export.mjs:14 core/export.mjs:33 core/export.mjs:69 core/export.mjs:28 core/export.mjs:31 core/export.mjs:36 core/export.mjs:42 core/export.mjs:50 core/export.mjs:56 core/export.mjs:58 core/export.mjs:66 core/export.mjs:79 core/export.mjs:83 core/export.mjs:86 core/fsguard.mjs:16 core/fsguard.mjs:20 core/fsguard.mjs:28 core/fsguard.mjs:48 core/render-frames.mjs:42 core/render-frames.mjs:84 core/render-frames.mjs:87 core/render-frames.mjs:88 core/render.mjs:5 core/render.mjs:31 core/render.mjs:30 core/scrub.mjs:34 core/scrub.mjs:43 core/scrub.mjs:46 core/scrub.mjs:74 core/setup.mjs:23 core/setup.mjs:34 core/setup.mjs:48 core/setup.mjs:5

🔑 Variables d’environnement (50)

🌐 Accès réseau (24)

Audité par: codex

Partager et citer ce rapport

Partagez le rapport d’évaluation versionné, le badge neutre, la carte intégrable et les citations. Skillstore présente les preuves sans décider si ce Skill est sûr.

Ouvrir le rapport versionné

Copier le lien du rapport

https://skillstore.io/skills/israelvicars-agent-transit/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Badge Markdown

[![Skillstore security assessment](https://skillstore.io/badges/skills/israelvicars-agent-transit/security.svg)](https://skillstore.io/skills/israelvicars-agent-transit?utm_source=security_passport_badge)

Badge HTML

<a href="https://skillstore.io/skills/israelvicars-agent-transit?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/israelvicars-agent-transit/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Carte intégrable

<iframe src="https://skillstore.io/embed/skills/israelvicars-agent-transit.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

Citations académiques (APA · BibTeX · CFF)

Citation APA

israelvicars. (2026). agent-transit security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/israelvicars-agent-transit/audits/1

Citation BibTeX

@techreport{israelvicars-israelvicars-agent-transit-2026,
  author = {israelvicars},
  title = {agent-transit security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/israelvicars-agent-transit/audits/1},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "agent-transit security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
  - name: "israelvicars"
date-released: "2026-10-22"
url: "https://skillstore.io/skills/israelvicars-agent-transit/audits/1"
identifiers:
  - type: other
    value: "skillstore:israelvicars-agent-transit:audit:1"
    description: "Skillstore immutable audit report identifier"

Score Skillstore

Pourquoi ce score Niveau de confiance des preuves: Moyen

Architecture

Maintenabilité

Contenu

Communauté

Conformité aux spécifications

Ce que vous pouvez créer

Investigate a failed coding session

Review tool order, slow operations, and denied calls in one saved local session.

Review a session with teammates

Prepare a scrubbed HTML map and inspect retained labels before sharing the artifact.

Prepare a synthetic workflow demonstration

Render a synthetic agent session into GIF or MP4 for a training presentation.

Essayez ces prompts

Map a saved session

Create a local HTML transit map from [session path] using [source format]. Keep argument previews disabled and do not upload anything.

Inspect the latest session failure

Find my latest Codex session for this workspace. Build a strict-share map and report the first failed call without exposing its payload.

Export a Cursor composer

List Cursor composers for [workspace]. After I select an identifier, export its metadata and build a local map with timing caveats.

Render and verify a session animation

Render [session path] as HTML, GIF, and MP4 with original wall-clock timing. Ask before downloads, verify outputs, and review retained data before sharing.

Bonnes pratiques

Keep original sessions local and inspect every generated artifact for confidential labels before sharing.
Use strict sharing, leave argument previews disabled, and treat scrubbing as partial protection.
Start with synthetic fixtures and approve dependency downloads before rendering sensitive sessions.

Éviter

Treating strict-share output as guaranteed free of confidential data.
Interpreting estimated or synthetic Cursor durations as exact execution times.
Sharing original logs or enabling argument previews without reviewing retained data.

Foire aux questions

Which coding agents are supported?

The adapters support Claude Code, Codex CLI, and experimental Cursor session formats.

Does session data leave my machine?

HTML generation is local. Rendering setup may download dependencies, and optional request-trace adapters contact configured services. Uploads require a separate sharing action.

Does strict sharing remove every secret?

No. It adds scrubbing, but query fragments and extra internal trace fields can remain. Review output before sharing.

What software is required?

HTML builds require Node.js 20 or later. Video rendering also requires Chromium and ffmpeg; setup may download missing components.

How accurate is Cursor timing?

Newer bubbles provide start timestamps, but end times use the next message as an upper bound. Older transcripts use uniform synthetic durations.

Can I export animations and check their rendering?

The bundled scripts produce GIF, MP4, and contact sheets. Verification compares repeated frames and checks departure timing and blank frames.