mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-07-22 16:25:58 +00:00
0f102fa4dc
* feat(browser): store full snapshots on truncation; make eval denylist opt-in Two harness fixes motivated by BU_Bench results where fixed-verb + lossy observation cost Hermes heavily vs code-driven browser agents: 1. Snapshot truncation no longer loses content. When a snapshot exceeds the 8000-char threshold, the complete accessibility tree is saved to cache/web (same truncate-and-store pattern as web_extract) and the truncated view / LLM summary includes the file path plus a ready-made read_file call. Element refs beyond the cut are recoverable without re-snapshotting. Stored copies are force-redacted and capped at 2MB; content-hash filenames dedupe repeated snapshots of the same page. 2. The browser_console(expression=...) sensitive-primitive denylist is now opt-in via browser.restrict_evaluate (default false). The names-based denylist blocked legitimate DOM extraction — any selector or expression containing 'fetch', 'cookie', 'input', etc. — which crippled the agent's only programmatic page-inspection path. The SSRF/private-URL egress guards in _browser_eval are independent of this policy and remain always-on. browser.allow_unsafe_evaluate keeps its meaning (bypass the denylist) for configs that already set it. * test: update None-guard test for stored-snapshot pointer in _extract_relevant_content test_normal_content_returned pinned the exact return value; the summary now carries a pointer to the stored full snapshot. Assert the summary passes through and the pointer is present instead. * feat(browser): align snapshot threshold with web_extract's 15k char budget SNAPSHOT_SUMMARIZE_THRESHOLD 8000 -> 15000, matching web_tools.DEFAULT_EXTRACT_CHAR_LIMIT so the snapshot and web_extract truncate-and-store paths give the model the same per-page budget. _truncate_snapshot's default max_chars now follows the constant. Invariant test added; docs (en+zh) and CLI tip updated. |
||
|---|---|---|
| .. | ||
| _category_.json | ||
| acp.md | ||
| api-server.md | ||
| batch-processing.md | ||
| browser.md | ||
| built-in-plugins.md | ||
| code-execution.md | ||
| codex-app-server-runtime.md | ||
| computer-use.md | ||
| context-files.md | ||
| context-references.md | ||
| credential-pools.md | ||
| cron.md | ||
| curator.md | ||
| delegation.md | ||
| deliverable-mode.md | ||
| extending-the-dashboard.md | ||
| fallback-providers.md | ||
| goals.md | ||
| honcho.md | ||
| hooks.md | ||
| image-generation.md | ||
| kanban-tutorial.md | ||
| kanban-worker-lanes.md | ||
| kanban.md | ||
| lsp.md | ||
| mcp.md | ||
| memory-providers.md | ||
| memory.md | ||
| mixture-of-agents.md | ||
| overview.md | ||
| personality.md | ||
| pets.md | ||
| plugins.md | ||
| provider-routing.md | ||
| skills.md | ||
| skins.md | ||
| spotify.md | ||
| subscription-proxy.md | ||
| tool-gateway.md | ||
| tool-search.md | ||
| tools.md | ||
| tts.md | ||
| vision.md | ||
| voice-mode.md | ||
| web-dashboard.md | ||
| web-search.md | ||
| x-search.md | ||