Skip to content

test(heartbeat): calibrate complete thin JSON scope budget #7614

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
loopx-agent merged 1 commit into from
Oct 28, 2026
Merged

test(heartbeat): calibrate complete thin JSON scope budget #7614

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 20 additions & 0 deletions docs/development/testing-and-quality.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1580,6 +1580,26 @@ not a real model's adoption, benchmark improvement or additional execution quota
载体从缺失到完整指引的首次变化给予 2400/3600/600 的有界额度；绝对上限、语义、
规模及后续增长检查不变。原失败保留；不据此声称模型采用、实验增益或额外执行权限。

The thin heartbeat's complete first-guard selection guidance also needs its
supported 320-character authored scope. At immutable main `b47bbc6c9`, the
unchanged public CLI fixture emits 4,243 JSON characters with the default scope
and 4,532 with the full supported scope; the latter fails the old 4,500 regression
ceiling. Keep that failure and all instruction text. Calibrate only this
surface's JSON ceiling to 4,600, leaving 68 characters of headroom at the supported
scope. The candidate must emit the same JSON and task body on that same fixture;
this is a budget decision, not an output reduction. The presentation calculation
remains 4,181 / 4,200 characters for that body (distinct from its raw 4,276-character
string). Markdown, line, growth, semantic and frozen differential/promotion
limits remain unchanged. No execution, spending, host loading or permission
limit changes.

thin 心跳保留完整的首次 guard 选择指引与受支持的 320 字符 scope。相同公开 CLI
夹具在固定主干上分别输出 4,243 / 4,532 个 JSON 字符，后者仍记录为旧 4,500
回归预算失败。仅将该 surface 的 JSON 预算校准为 4,600，留 68 字符余量；候选
必须保持同一夹具的 JSON 与正文不变，不宣称压缩收益。正文展示口径仍为
4,181 / 4,200，与 raw string 的 4,276 字符分列；Markdown、行数、增长、语义、
冻结差分与晋级标准、执行额度和权限均保持原边界。

The latest qualification pins true main `e47e4507a` and merged candidate
`3955d1c8b` on the same public fixture, with explicit per-probe source imports.
It retains 1/36/18 Todos, 1/12/12 history records, complete command routes,
Expand Down
13 changes: 7 additions & 6 deletions loopx/control_plane/testing/cli_output_budget.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -312,12 +312,13 @@ class CliOutputCommandClassification:
),
markdown_anchor="# Heartbeat Automation Prompt",
max_chars={
# Matched Linux output: JSON 3,649 -> 4,172 and Markdown
# 5,553 -> 6,076. The supported 320-char scope reaches 4,461
# JSON chars. Keep the full outcome guidance and Markdown guard.
"small": {"json": 4_500, "markdown": 6_200},
"crowded": {"json": 4_500, "markdown": 6_200},
"multi_agent": {"json": 4_500, "markdown": 6_200},
# The complete first-guard selection guidance and supported
# 320-char scope emit 4,532 JSON chars at immutable b47bbc6c9.
# Retain the old 4,500 failure; allow 68 chars of JSON headroom.
# Body presentation stays 4,200; Markdown/line guards are unchanged.
"small": {"json": 4_600, "markdown": 6_200},
"crowded": {"json": 4_600, "markdown": 6_200},
"multi_agent": {"json": 4_600, "markdown": 6_200},
},
max_lines={
"small": {"json": 18, "markdown": 72},
Expand Down
Loading