legion-llm 0.7.6 → 0.7.8

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
checksums.yaml CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA256:
3
- metadata.gz: 2856c4fbdb896cc947ca9e64e1f16240db499ce49edb92d1a775009de9405a7b
4
- data.tar.gz: dde486aca3ebfefec431e4ac4578db846a2fcd597a118657ba44cfca05245cc6
3
+ metadata.gz: 95918a13b81cc27203ad7eaccf52b06b7989735cb669febc7e94abd8c8c5452c
4
+ data.tar.gz: a7887cd39bbeda75313f37c134ef8461d7d6bd240be811dbf77e91f64343fc30
5
5
  SHA512:
6
- metadata.gz: 96435d5a198f879241b1c04c2c17cd0caf2afd2ae3751c2efd6c416e19c3513bbfe3c3612289a4014fed6f461e9c801451e6fd767fa23c62075cd908cdb6f676
7
- data.tar.gz: 4978d31e8be17eadb107ac10481f1089ed32bb5403e60e1f5320510b6d93c7f73a1880d063ac55e05144dab64998839f7280c8e9271378adceaf08b41b4a35d8
6
+ metadata.gz: e389df67e9855d573a195948123e6aae6760da6e461d54148edd0cd1faa8535445c744ce02b3512e89fd7603197370f4ea8d84bb9dd82588d3e545dea308bc2b
7
+ data.tar.gz: 965932e45bf5905b77ba667d43d24e70380a203b048f01daaa17cc72c835a952fd9ee474a78a868bd8fef538b1dca9843d599f13249cb88d7c67eca6b53fd64b
data/CHANGELOG.md CHANGED
@@ -2,6 +2,19 @@
2
2
 
3
3
  ## [Unreleased]
4
4
 
5
+ ## [0.7.8] - 2026-04-17
6
+ ### Fixed
7
+ - Guard `Embeddings.generate` against providers that don't support embeddings before calling `RubyLLM.embed` — prevents noisy `NoMethodError: undefined method 'render_embedding_payload'` warns when Bedrock (or Anthropic) is the active provider
8
+
9
+ ## [0.7.7] - 2026-04-15
10
+
11
+ ### Added
12
+ - Sticky runner tool injection: deferred tool runners stay injected for N human turns (trigger tier) or N deferred tool executions (execution tier) after activity, preventing the LLM from falling back to `legion_do` on follow-up messages
13
+ - Tool call history: every tool call is appended to a per-conversation enrichment block so the LLM can reference prior results (e.g. "the issue I created was #142")
14
+ - `Steps::StickyRunners`, `Steps::ToolHistory`, `Steps::StickyPersist` pipeline steps
15
+ - `ConversationStore#read_sticky_state` / `#write_sticky_state` for in-memory per-conversation state
16
+ - `Steps::StickyHelpers` shared settings helpers
17
+
5
18
  ## [0.7.6] - 2026-04-14
6
19
 
7
20
  ### Added