Commit graph

7 commits

Author SHA1 Message Date
aaronagent
1909877e6e fix: cap image download size at 50 MB, validate tool call parser fields
vision_tools.py: _download_image() loads the full HTTP response body into
memory via response.content (line 190) with no Content-Length check and no
max file size limit.  An attacker-hosted multi-gigabyte file causes OOM.
Add a 50 MB hard cap: check Content-Length header before download, and
verify actual body size before writing to disk.

hermes_parser.py: tc_data["name"] at line 57 raises KeyError when the LLM
outputs a tool call JSON without a "name" field.  The outer except catches
it silently, causing the entire tool call to be lost with zero diagnostics.
Add "name" field validation before constructing the ChatCompletionMessage.

mistral_parser.py: tc["name"] at line 101 has the same KeyError issue in
the pre-v11 format path.  The fallback decoder (line 112) already checks
"name" correctly, but the primary path does not.  Add validation to match.

Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>
2026-04-11 02:03:20 -07:00
Himess
5663980015
fix(mistral-parser): handle nested JSON in fallback extraction 2026-03-21 09:41:17 -07:00
Emmett
26bedf973b
fix: support multiple parallel tool calls in DeepSeek V3 parser (#989)
- Refactored regex pattern to handle varied whitespace and newlines for better robustness.
- Replaced logic to iterate through all tool call blocks using finditer instead of stopping at the first match.
- Ensured full extraction of multiple tool calls for complex agentic workflows.
- Added error logging for failed parsing attempts.
2026-03-15 03:55:24 +01:00
Teknium
2ff03ebafe
fix: use non-greedy regex in DeepSeek V3 parser for multi-tool calls (#1300)
The greedy `.*` captures with `re.DOTALL` cause `findall()` to merge
multiple tool calls into a single match — silently dropping all but the
last tool call. Switching to `.*?` (non-greedy) fixes extraction when
models return multiple tool calls in one response.

Adds test coverage for the DeepSeek V3 parser including a multi-tool
call regression test.

Co-authored-by: Himess <semihcvlk53@gmail.com>
2026-03-14 06:19:28 -07:00
teknium1
dc55f493be fix: add missing re.DOTALL to DeepSeek V3.1 parser (same bug as V3)
The V3.1 parser had the same issue — .*? without re.DOTALL fails to
match multi-line JSON arguments. Found during review of PR #444.
2026-03-06 04:41:47 -08:00
PercyDikec
e36c8cd49a
fix: add missing re.DOTALL flag to DeepSeek V3 tool call parser 2026-03-05 20:32:38 +03:00
teknium
07b615e96e Add support for Atropos Agentic RL environments (requires branch tool_call_support in Atropos atm)
- Added new environments for reinforcement learning, including `HermesSweEnv` for software engineering tasks and `TerminalTestEnv` for inline testing.
- Introduced `ToolContext` for unrestricted access to tools during reward computation.
- Updated `.gitignore` to exclude `wandb/` directory.
- Enhanced `README.md` with detailed architecture and usage instructions for Atropos environments.
- Added configuration files for SWE and terminal test environments to streamline setup.
- Removed unnecessary compiled Python files from `__pycache__`.
2026-02-07 09:17:16 +00:00