Releases
v0.6.13
Compare
Sorry, something went wrong.
No results found
@workglow/ai
Features
providers
carry reasoning, tool calls and images between rounds on local runtimes
carry reasoning and tool-result images between rounds on cloud APIs
openai,anthropic
hand a model its own turn back each round
ai
AgentTask deadlines, time-left notices, review, reasoning and native-turn storage
Bug Fixes
ai
fail closed when a round reports no usage under maxCostUsd
do not duplicate a retried AgentTask round's partial text
name the field a nullable section's submission got wrong
@workglow/test
Features
providers
carry reasoning, tool calls and images between rounds on local runtimes
carry reasoning and tool-result images between rounds on cloud APIs
openai,anthropic
hand a model its own turn back each round
ai
AgentTask deadlines, time-left notices, review, reasoning and native-turn storage
Bug Fixes
truncate
improve handling of newline allowance when truncating text
anthropic
add test for handling disabled thinking in native turn
tasks
one credential-name pattern for URL redaction and body scrubbing
read an error page's title without a backtracking pattern
redact credentials in the URL quoted by fetch error messages
scrub credentials and neutralise remote text in HTTP fetch errors
ai
fail closed when a round reports no usage under maxCostUsd
do not duplicate a retried AgentTask round's partial text
name the field a nullable section's submission got wrong
Chores
Updated Dependencies
@workglow/tasks
Bug Fixes
tasks
one credential-name pattern for URL redaction and body scrubbing
read an error page's title without a backtracking pattern
redact credentials in the URL quoted by fetch error messages
scrub credentials and neutralise remote text in HTTP fetch errors
Chores
Updated Dependencies
@workglow/openai
Features
openai,anthropic
hand a model its own turn back each round
@workglow/llamacpp-server
Features
providers
carry reasoning, tool calls and images between rounds on local runtimes
@workglow/ollama
Features
providers
carry reasoning, tool calls and images between rounds on local runtimes
@workglow/node-llama-cpp
Features
providers
carry reasoning, tool calls and images between rounds on local runtimes
@workglow/anthropic
Features
openai,anthropic
hand a model its own turn back each round
Bug Fixes
anthropic
add test for handling disabled thinking in native turn
agent-eval
enhance time management and credential scrubbing
@workglow/google-gemini
Features
providers
carry reasoning and tool-result images between rounds on cloud APIs
@workglow/xai
Features
providers
carry reasoning and tool-result images between rounds on cloud APIs
@workglow/deepseek
Features
providers
carry reasoning and tool-result images between rounds on cloud APIs
@workglow/huggingface-transformers
Features
providers
carry reasoning, tool calls and images between rounds on local runtimes
@workglow/tf-mediapipe
Bug Fixes
@workglow/chrome-ai
Features
providers
carry reasoning, tool calls and images between rounds on local runtimes
@workglow/openrouter
Features
providers
carry reasoning and tool-result images between rounds on cloud APIs
@workglow/huggingface-inference
Features
providers
carry reasoning and tool-result images between rounds on cloud APIs
@workglow/cactus
Features
providers
carry reasoning, tool calls and images between rounds on local runtimes
Bug Fixes
truncate
improve handling of newline allowance when truncating text
@workglow/eval
Chores
Updated Dependencies
@workglow/cli
Features
cli,web
draw a task's reasoning apart from its answer
Chores
Updated Dependencies
ink: ^8.0.0
smol-toml: ^1.9.1
preact: ^11.0.1
@workglow/web
Features
cli,web
draw a task's reasoning apart from its answer
Chores
Updated Dependencies
@codemirror/view: 6.43.14
@codemirror/language: 6.13.1
@lezer/lr: 1.4.11
react-resizable-panels: ^4.14.5
@vitejs/plugin-react: ^6.1.2
jsdom: ^30.1.2
vite: ^8.3.4
@workglow/agent-eval
Features
agent-eval
compare AgentTask with opencode and pi on public benchmarks
Bug Fixes
truncate
improve handling of newline allowance when truncating text
agent-eval
enhance time management and credential scrubbing
You can’t perform that action at this time.