Install verification: runtime verified (L5) · tested on dsh 0.2.1-rc.2 · last checked Install verification details →
What it does
Embed chat.deepseek.com web chat in DSH and transfer the conversation into a workspace session for DeepSeek Harness.
Installation
dsh plugin --profile web add github:naiIong/dsh-webchat-bridgeInstall method: GitHub · runtime verified (L5)
Verification & Compatibility
Grouped by plugin version: each card lists the dsh versions actually tested against that artifact version and their install verdicts (L5 wins). Untested combinations are omitted — absence means untested, never assumed compatible. L1–L3 are a plugin-level static archive (independent of the dsh version — the same verdict on every dsh row): they sit in each row's level grid next to L4/L5, marked "static", while their evidence is stored once per plugin inside that row's "checks & evidence" instead of being duplicated per version.
External credentials (plugin-level): none detected
○ The author declares no dsh version requirement — install is not blocked by version.
| plugin \ dsh | dsh 0.2.1-rc.2 current |
|---|---|
| 59cd5c | ✓ |
59cd5cgithub✓ 1 runtime-verified2026-10-19
https://github.com/naiIong/dsh-webchat-bridge
✓dsh 0.2.1-rc.2current2026-10-19L5 · runtime verified
- • Node.js: v24.14.0
- • DSH: 0.2.1-rc.2
- • Plugin artifact: github · https://github.com/naiIong/dsh-webchat-bridge (version unstated)
ID: dsh-webchat-bridge@59cd5c@dsh0.2.1-rc.2 · profile: l4-sandbox
Issued: 2026-10-19 · expires: 2026-10-26
View checks & evidence
active in loader tree (150 entries)
L1–L3 are plugin-level static conclusions (baseline, identical across dsh versions): evidence is stored once per plugin.
Author badges
Embed the official badges in your README to showcase security & install-test status:
`[](https://www.dsh.so/artifact/dsh-webchat-bridge/)``[](https://www.dsh.so/artifact/dsh-webchat-bridge/)`Security Report
Two automated layers, neither a manual review: a static scan (per-plugin rule checks) and a model-driven dynamic security audit that covers only the subset static-flagged at warn or above and never executes plugin code.
Static low, dynamic medium — stricter wins → overall medium.
v0.1.0dsh manifest0.1.0git2026-10-210.1.0same versionclean · gitVet is an AST static scan (npm artifact or git source).
show static findings (1)
- criticalnewFunction · code-exec
lib/index.js:232schema.callback = new Function("return " + schema.callback)();
show 3 audit findings
- mediumState-changing HTTP endpoints have no auth, origin or CSRF check; access control is fully delegated to the host
lib/index.jsAny client able to reach the DSH web server can drive the privileged host half: `send` types attacker-supplied text into the signed-in chat.deepseek.com composer and presses Enter (data leaves the machine under the user's account), `transfer` writes attacker text into new or existing workspace sessions, and `import`/`rename`/`remove`/`clear`/`toggles` mutate plugin state. Because `readJson` ignores Content-Type, a cross-site page can issue a simple (non-preflighted) POST with a text/plain JSON body at these URLs and the side effect will still execute even though the response is unreadable.
- lowDeepSeek bearer token is scraped from page requests and persisted in plaintext although nothing reads it
lib/index.jsA live account credential is copied out of browser traffic and kept indefinitely as plaintext JSON next to the profile; `readToken`/`hasToken` are only used to report `tokenStored`, so the secret is retained without any functional need. Exposure is limited to principals that can read the user's data dir (same-user processes, backups, mis-set permissions), but it widens the blast radius beyond the browser profile.
- lowUntrusted web-chat or imported text is written into sessions with user-role authority
lib/index.jsContent that originates from chat.deepseek.com model output or from an imported file is promoted into a DSH session as a `user`-sourced message. The downstream coding agent cannot distinguish it from instructions the human actually typed, so injected text in the transferred transcript can steer tool use with user-level authority. The action is user-initiated and the preamble text states the provenance, which keeps this low.
Privileged browser-automation and session-write HTTP routes rely wholly on host auth and persist a captured DeepSeek bearer token; the static-scan critical (vendored new Function) is unreachable.
Rule: the dynamic audit can only raise risk, never lower it (raise-only, never an exemption); a wrong raise is withdrawn by human review and the tier falls back to the static verdict. Overall verdict = the stricter of the two layers.
Disclaimer: the static scan and the dynamic security audit are both automated analysis, not a security guarantee. Always review what you install.
Activity
Last commit 2026-10-18 · activity: Active
• Repo created: 2026-10-18
• Stars: ★ 0 · Forks: 0
• Health: Active — committed within last 30 days
Source
Author README
View on GitHub ↗The author's original README from the repository — fetched live, unedited (the English page loads the default README).
