AGENT SKILL

agent-reach-browser-bridge

Scaffold, configure, and format multi-platform channel routing, credential scrubbing, and unified Schema v1 result envelopes for AI agents. Envelope formatting only. Do not use for generic SQL querying or raw binary network sniffing.

License 许可
MIT
最近更新
2026-11-15
Stars
★ 0
Forks
0

01 前言 Front Matter

name
agent-reach-browser-bridge
description
Scaffold, configure, and format multi-platform channel routing, credential scrubbing, and unified Schema v1 result envelopes for AI agents. Envelope formatting only. Do not use for generic SQL querying or raw binary network sniffing.

02 资源文件 3

03 SKILL.md 正文

Agent Reach Browser Bridge: Architecture & Integration

The agent-reach-browser-bridge skill provides structured domain capabilities for orchestrating multi-platform web extraction, walled-garden content retrieval (YouTube, GitHub, Reddit, Twitter, Bilibili, Xiaohongshu, Boss直聘, V2EX), dedicated anti-fingerprint CloakBrowser sessions over Chrome DevTools Protocol (CDP), and normalized Schema v1 result envelopes for autonomous AI coding agents.

See CARD.md for the companion summary card, stage progression table, and invariants checklist. Consult config.default.yaml for baseline operational budgets and timeout thresholds.


The Three Foundational Pillars

1. Unified Schema v1 Result Envelope & Error Taxonomy

Standardizing all scraping and extraction outputs into an immutable {schema_version, ok, channel, action, backend, items, meta, error} JSON envelope. Categorizes failures into deterministic final codes (FINAL_CODES: NEED_LOGIN, RATE_LIMITED, PLATFORM_BLOCKED, NOT_FOUND, BAD_INPUT) versus transient retryable codes (TIMEOUT, UPSTREAM_BROKEN, NEED_SETUP), allowing agent reasoning loops to branch reliably without brittle DOM parsing.

2. Deterministic Multi-Tier Backend Fallback Cascades

Decoupling user actions from fragile single-scraper implementations through ordered backend cascades:

  • Tier 1 (Public Web API): Zero-credential low-latency JSON endpoints.
  • Tier 2 (Stealth Browser): Local CloakBrowser automation over loopback CDP (127.0.0.1).
  • Tier 3 (Native Tool / CLI): Authenticated third-party command-line utilities. Applies adaptive circuit breaking (BACKEND_PAUSE_SECONDS) to pause failing backends and prevent agent compute thrashing.

3. Non-Invasive Session Sandboxing & Credential Redaction

Enforcing strict filesystem isolation by redirecting third-party CLI HOME and USERPROFILE environments into private per-tool directories (~/.agent-reach/sessions/<tool>/). Validates symlink safety and automatically scrubs inline HTTP basic auth credentials (scrub_url_credentials) before surfacing outputs to the agent context.


Operational Workflow

[Agent Tool Request] ──► [URL / Channel Matcher (can_handle)]
                                   │
                                   ▼
                   [Layered Backend Fallback Cascade]
                 ┌─────────────────┼─────────────────┐
                 ▼                 ▼                 ▼
          [1. Public API]   [2. CloakBrowser]  [3. Native CLI]
          (Zero-Auth REST)   (CDP 127.0.0.1)   (Private $HOME)
                 └─────────────────┬─────────────────┘
                                   │
                                   ▼
                      [Credential Scrubber Engine]
                                   │
                                   ▼
                   [Unified Schema v1 JSON Envelope]

Anti-Patterns

  • Raw HTML Extraction Output — Returning unparsed HTML or noisy terminal stdout directly to LLM context instead of Schema v1 JSON envelopes.
  • Host User Home Contamination — Allowing third-party scraping utilities to dump login cookies into the default $HOME instead of private sandboxes.
  • Unbounded Throttled Retries — Continuing to retry rate-limited or login-gated endpoints without honoring FINAL_CODES circuit breakers.
  • Zombie Browser Leaks — Failing to shut down idle browser processes after completion, resulting in stranded Chromium handles.
  • Plaintext Credential Leakage — Permitting user:password secrets in URLs to pass unscrubbed into model trajectories.

04 同仓库更多 Skills

SpectreDeath
SpectreDeath/Brain-Harness

ui-behavior-primitive-architect

Architect, audit, and verify zero-dependency headless UI behavior primitives, WebMCP agent bridges, WAI-ARIA/SSR/RTL tri-modal checks, and framework expert routing. Do not use for static CSS theme styling or backend database migrations.

★ 0
SpectreDeath
SpectreDeath/Brain-Harness

trustgraph-ontology-governor

Govern knowledge graphs, SHACL-AF triple policies, OntologyRAG subclass validation, PROV-O dual named graphs, and Laya entropy decision scoring. Do not use for relational SQL queries or generic web scraping.

★ 0
SpectreDeath
SpectreDeath/Brain-Harness

shim-harness-governor

Govern universal CLI agent loops, 4-tier tool permissions, ancestor-symlink workspace sandboxes, deterministic-first command routing, collision-safe MCP namespacing, and layered skill loading. Do not use for static UI styling or external SQL analytics.

★ 0
SpectreDeath
SpectreDeath/Brain-Harness

graphiti-zep-memory-governor

Govern bi-temporal knowledge graphs with Graphiti future-edge clamping, MinHash/LSH entity dedup, RRF/MMR reranking, and Zep typed ontologies. Do not use for static SQL joins or generic web crawling.

★ 0
SpectreDeath
SpectreDeath/Brain-Harness

agora-mcp-workbench-architect

Architect, audit, and orchestrate Agora Workbench MCP servers, loopback sidecars, unified state graphs, gateway/dispatcher connectors, and policy-enforced data lakes. Do not use for static frontend CSS styling or non-MCP desktop GUI automation.

★ 0
SpectreDeath
SpectreDeath/Brain-Harness

skill-eval-assurance-runner

Evaluate and assure agent skills using small-N Wilson confidence intervals, unbiased pass^k reliability, deterministic trajectory graders, preflight boundary honesty, and sealed digest verification. Do not use for subjective prose review or unconstrained production deployment.

★ 0