agenr 0.8.25 → 0.8.27

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.
package/CHANGELOG.md CHANGED
@@ -1,5 +1,47 @@
1
1
  # Changelog
2
2
 
3
+ ## [0.8.27] - 2026-02-24
4
+
5
+ ### Changed
6
+ - Add stderr debug probes to openclaw plugin to diagnose before_reset hook dispatch issue
7
+ - Probes: register() entry, hook registrations, session_start handler, before_reset handler entry and guard points
8
+
9
+ ## [0.8.26] - 2026-02-23
10
+
11
+ ### Added
12
+ - feat(openclaw-plugin): LLM-summarized multi-session handoff entry at
13
+ before_reset (#199). Builds a merged transcript from the current session
14
+ (via event.messages) and the most recent prior .reset.* file (if under
15
+ 24h old), labeled with timestamps and surface (webchat/telegram/etc.)
16
+ from sessions.json. Summarizes via the configured LLM (from agenr config,
17
+ default gpt-4.1-nano) into a structured four-section handoff. Falls back
18
+ to raw text extraction on any failure. Handler is now properly async with
19
+ awaited store call.
20
+
21
+ ### Changed
22
+ - fix(openclaw-plugin): before_reset handoff store now uses a two-phase flow to
23
+ avoid race windows with detached hook execution. Phase 1 stores fallback
24
+ exchange text immediately at importance 9, then Phase 2 asynchronously upgrades
25
+ to an importance 10 LLM summary when available.
26
+ - fix(openclaw-plugin): when the LLM upgrade succeeds, fallback handoff entries
27
+ are looked up and retired (subject/content/tag match) before storing the
28
+ upgraded summary, preventing stale fallback carryover.
29
+ - fix(openclaw-plugin): prior reset session surface lookup now maps
30
+ `*.jsonl.reset.*` files back to base `*.jsonl` paths via getBaseSessionPath,
31
+ and unknown surface fallback now uses "prior session" to improve prompt context.
32
+ - fix(openclaw-plugin): capTranscriptLength now enforces a hard length cap even
33
+ when the current session alone exceeds 8000 chars.
34
+ - chore(openclaw-plugin): added before_reset debug logs for missing sessionFile,
35
+ missing apiKey, and pre-LLM invocation traceability.
36
+
37
+ ### Tests
38
+ - test(openclaw-plugin): added coverage for getBaseSessionPath, reset-path surface
39
+ resolution, capTranscriptLength edge cases (prior-only overflow, current-only
40
+ overflow, under-cap passthrough), missing-apiKey debug behavior, budget tail
41
+ slicing assertions, and buildMergedTranscript ordering.
42
+ - test(openclaw-plugin): updated before_reset integration coverage for two-phase
43
+ fallback-plus-upgrade storage behavior and no-sessionFile debug path.
44
+
3
45
  ## [0.8.25] - 2026-02-23
4
46
 
5
47
  ### Changed