hermes-agent/agent
Andre Kurait a8e89cbbf6
fix(bedrock): resolve context length via static table before custom-endpoint probe
## Problem

`get_model_context_length()` in `agent/model_metadata.py` had a resolution
order bug that caused every Bedrock model to fall back to the 128K default
context length instead of reaching the static Bedrock table (200K for
Claude, etc.).

The root cause: `bedrock-runtime.<region>.amazonaws.com` is not listed in
`_URL_TO_PROVIDER`, so `_is_known_provider_base_url()` returned False.
The resolution order then ran the custom-endpoint probe (step 2) *before*
the Bedrock branch (step 4b), which:

  1. Treated Bedrock as a custom endpoint (via `_is_custom_endpoint`).
  2. Called `fetch_endpoint_model_metadata()` → `GET /models` on the
     bedrock-runtime URL (Bedrock doesn't serve this shape).
  3. Fell through to `return DEFAULT_FALLBACK_CONTEXT` (128K) at the
     "probe-down" branch — never reaching the Bedrock static table.

Result: users on Bedrock saw 128K context for Claude models that
actually support 200K on Bedrock, causing premature auto-compression.

## Fix

Promote the Bedrock branch from step 4b to step 1b, so it runs *before*
the custom-endpoint probe at step 2. The static table in
`bedrock_adapter.py::get_bedrock_context_length()` is the authoritative
source for Bedrock (the ListFoundationModels API doesn't expose context
window sizes), so there's no reason to probe `/models` first.

The original step 4b is replaced with a one-line breadcrumb comment
pointing to the new location, to make the resolution-order docstring
accurate.

## Changes

- `agent/model_metadata.py`
  - Add step 1b: Bedrock static-table branch (unchanged predicate, moved).
  - Remove dead step 4b block, replace with breadcrumb comment.
  - Update resolution-order docstring to include step 1b.

- `tests/agent/test_model_metadata.py`
  - New `TestBedrockContextResolution` class (3 tests):
    - `test_bedrock_provider_returns_static_table_before_probe`:
      confirms `provider="bedrock"` hits the static table and does NOT
      call `fetch_endpoint_model_metadata` (regression guard).
    - `test_bedrock_url_without_provider_hint`: confirms the
      `bedrock-runtime.*.amazonaws.com` host match works without an
      explicit `provider=` hint.
    - `test_non_bedrock_url_still_probes`: confirms the probe still
      fires for genuinely-custom endpoints (no over-reach).

## Testing

  pytest tests/agent/test_model_metadata.py -q
  # 83 passed in 1.95s (3 new + 80 existing)

## Risk

Very low.

- Predicate is identical to the original step 4b — no behaviour change
  for non-Bedrock paths.
- Original step 4b was dead code for the user-facing case (always hit
  the 128K fallback first), so removing it cannot regress behaviour.
- Bedrock path now short-circuits before any network I/O — faster too.
- `ImportError` fall-through preserved so users without `boto3`
  installed are unaffected.

## Related

- This is a prerequisite for accurate context-window accounting on
  Bedrock — the fix for #14710 (stale-connection client eviction)
  depends on correct context sizing to know when to compress.

Signed-off-by: Andre Kurait <andrekurait@gmail.com>
2026-04-23 20:33:09 +00:00
..
transports fix: add extra_content property to ToolCall for Gemini thought_signature (#14488) 2026-04-23 23:45:07 +05:30
__init__.py Refactor Terminal and AIAgent cleanup 2026-02-21 22:31:43 -08:00
account_usage.py feat(account-usage): add per-provider account limits module 2026-04-21 01:56:35 -07:00
anthropic_adapter.py refactor: remove _nr_to_assistant_message shim + fix flush_memories guard 2026-04-23 02:30:05 -07:00
auxiliary_client.py feat: add Xiaomi MiMo v2.5-pro and v2.5 model support (#14635) 2026-04-23 10:06:25 -07:00
bedrock_adapter.py feat: native AWS Bedrock provider via Converse API 2026-04-15 16:17:17 -07:00
codex_responses_adapter.py refactor: extract codex_responses logic into dedicated adapter 2026-04-20 11:53:17 -07:00
context_compressor.py fix: pass correct arguments in summary model fallback retry 2026-04-22 17:57:13 -07:00
context_engine.py refactor: remove dead code — 1,784 lines across 77 files (#9180) 2026-04-13 16:32:04 -07:00
context_references.py fix(agent): fall back when rg is blocked for @folder references 2026-04-20 01:56:41 -07:00
copilot_acp_client.py fix(security): apply file safety to copilot acp fs 2026-04-21 01:31:58 -07:00
credential_pool.py fix(auth): unify credential source removal — every source sticks (#13427) 2026-04-21 01:52:49 -07:00
credential_sources.py fix(auth): unify credential source removal — every source sticks (#13427) 2026-04-21 01:52:49 -07:00
display.py fix(display): render <missing old_text> in memory previews instead of empty quotes (#12852) 2026-04-19 22:45:47 -07:00
error_classifier.py fix(error_classifier): retry mid-stream SSL/TLS alert errors as transport 2026-04-22 17:44:50 -07:00
file_safety.py fix(security): apply file safety to copilot acp fs 2026-04-21 01:31:58 -07:00
gemini_cloudcode_adapter.py refactor: remove redundant local imports already available at module level 2026-04-21 00:50:58 -07:00
gemini_native_adapter.py refactor: remove redundant local imports already available at module level 2026-04-21 00:50:58 -07:00
gemini_schema.py fix(gemini): sanitize tool schemas for Google providers 2026-04-20 00:26:18 -07:00
google_code_assist.py fix(gemini-cli): surface MODEL_CAPACITY_EXHAUSTED cleanly + drop retired gemma-4-26b (#11833) 2026-04-17 15:34:12 -07:00
google_oauth.py feat(gemini): add Google Gemini CLI OAuth provider via Cloud Code Assist (free + paid tiers) (#11270) 2026-04-16 16:49:00 -07:00
image_gen_provider.py feat(plugins): pluggable image_gen backends + OpenAI provider (#13799) 2026-04-21 21:30:10 -07:00
image_gen_registry.py feat(plugins): pluggable image_gen backends + OpenAI provider (#13799) 2026-04-21 21:30:10 -07:00
insights.py Merge branch 'main' into feat/dashboard-skill-analytics 2026-04-20 05:25:49 -07:00
manual_compression_feedback.py fix(gateway): make manual compression feedback truthful 2026-04-10 21:16:53 -07:00
memory_manager.py feat(honcho): context injection overhaul, 5-tool surface, cost safety, session isolation (#10619) 2026-04-15 19:12:19 -07:00
memory_provider.py refactor(memory): drop on_session_reset — commit-only is enough 2026-04-15 11:28:45 -07:00
model_metadata.py fix(bedrock): resolve context length via static table before custom-endpoint probe 2026-04-23 20:33:09 +00:00
models_dev.py fix: normalize provider in list_provider_models to support aliases 2026-04-23 01:59:20 -07:00
nous_rate_guard.py fix: Nous Portal rate limit guard — prevent retry amplification (#10568) 2026-04-15 16:31:48 -07:00
prompt_builder.py feat(agent): add PLATFORM_HINTS for matrix, mattermost, and feishu (#14428) 2026-04-23 12:50:22 +05:30
prompt_caching.py fix(prompt-caching): skip top-level cache_control on role:tool for OpenRouter 2026-03-21 16:54:43 -07:00
rate_limit_tracker.py refactor: remove dead code — 1,784 lines across 77 files (#9180) 2026-04-13 16:32:04 -07:00
redact.py feat: replace kimi-k2.5 with kimi-k2.6 on OpenRouter and Nous Portal (#13148) 2026-04-20 11:49:54 -07:00
retry_utils.py feat(agent): add jittered retry backoff 2026-04-08 00:41:36 -07:00
shell_hooks.py feat: shell hooks — wire shell scripts as Hermes hook callbacks 2026-04-20 20:53:51 -07:00
skill_commands.py feat(skills+terminal): make bundled skill scripts runnable out of the box (#13384) 2026-04-21 00:39:19 -07:00
skill_utils.py fix(skills): follow symlinks in iter_skill_index_files 2026-04-22 17:43:30 -07:00
subdirectory_hints.py fix(agent): catch PermissionError in subdirectory hint discovery 2026-04-09 03:10:30 -07:00
title_generator.py fix: increase max_tokens for GLM 5.1 reasoning headroom 2026-04-22 18:44:07 -07:00
trajectory.py Refactor Terminal and AIAgent cleanup 2026-02-21 22:31:43 -08:00
usage_pricing.py fix(usage): read top-level Anthropic cache fields from OAI-compatible proxies 2026-04-22 17:40:49 -07:00