Card: correct Honest Limits after stability reruns (self-ref was temp-noise, not a failure); v0.1 stays recommended over v0.2

README.md

@@ -109,27 +109,31 @@ over your system message or tool calls.
 ## Honest limits — where it actually breaks
 
 We pushed it with a **hard trap-suite** (10 tasks built specifically to make a strong 4B fail — 8-hop
-dependent chains, an RPN evaluator, byte-exact files, self-referential puzzles). smeagle scored **6/10**,
-and *what it missed is the useful part* — double-check it on these:
+dependent chains, an RPN evaluator, byte-exact files, self-referential puzzles) and re-ran the flaky ones
+several times so we'd report *reliable* behaviour, not a lucky single draw. The genuine limits — double-check
+it on these:
 
 - **Byte-exact / multi-line formatting.** Told to write four exact lines with trailing spaces and no final
-  newline, it collapsed them onto one line. It'll *normalize* formatting — if you need bytes exactly, verify.
-- **Self-referential / fixed-point reasoning.** "Write a sentence that says how many words it has" → it wrote
-  "…has 1 word" (the true answer is 5). Puzzles whose answer depends on the answer trip it up.
+  newline, it collapsed them onto one line. It *normalizes* formatting — if you need bytes exactly, verify.
 - **Operand-order-sensitive code.** In a generated RPN evaluator it got the subtract/divide order wrong
   (`a op b` vs `b op a`) — the classic bug. Review order-sensitive logic it writes.
-- **Long single-file running state.** Overwriting one file with a running value across many steps, it can
-  lose the thread. Give it distinct filenames or fewer hops when precision matters.
+- **Long single-file running state.** Overwriting one file with a running value across many steps, it loses
+  the thread (0/3 in our reruns). Give it distinct filenames or fewer hops when precision matters.
+- **Temperature-sensitive on one-shot precision.** Some exact/self-referential tasks it gets right *most*
+  of the time but not every time (e.g. "write a sentence stating its own word count" — ~2 in 3 at default
+  temp 0.6). If you need it reliable in one shot, lower the temperature or let it retry.
 
-What it *does* do well (6/10 of the same brutal suite): long dependent tool-chains, multi-step error
-recovery (writing broken code, reading the error, fixing it), building larger working modules with correct
-logic, recognizing a tool's limits, and hitting strict output templates.
+What it *does* do well: the full agentic suite (16/16 — single-tool → multi-tool chains → file ops →
+code-gen+verify → error recovery → constrained output), plus long dependent tool-chains, building larger
+working modules with correct logic, and recognizing a tool's limits.
 
 - It's **4B** — a frontier model beats it on hard long-horizon work; smeagle's job is to be the *capable
   little one that fits*, not to out-muscle models 100× its size. And it's a **specialist** (agentic/terminal/
   SWE), not a generalist oracle.
-- v0.1. A v0.2 exists (trades a little SWE loss for agentic gains); we benchmark in the open and are still
-  deciding which serves you best. If you have a view, tell us.
+- **Why v0.1 and not v0.2?** We built a v0.2 and benchmarked both, in the open, on all 26 tasks with reruns.
+  v0.2's apparent edge turned out to be temperature noise once repeated — no *reliable* agentic gain — and it
+  regressed slightly on held-out SWE loss. So v0.1 stays the recommended release. That's the honest call, and
+  we'll only ship a v0.2 when it clearly earns it.
 
 ## License & rules — free for almost everyone