Conversation
Collaborator
Author
There was a problem hiding this comment.
Pull request overview
Removes the WeChat group announcement and embedded QR-code block from both the English and Chinese READMEs, keeping the project README content focused and avoiding references to external join links.
Changes:
- Delete the “WeChat group” changelog bullet from
README.mdandREADME.zh-CN.md. - Remove the centered HTML block that embedded the WeChat QR code image and its note in both READMEs.
Reviewed changes
Copilot reviewed 2 out of 3 changed files in this pull request and generated no comments.
| File | Description |
|---|---|
| README.zh-CN.md | Removes the WeChat group join note and QR-code embed from the Chinese README timeline section. |
| README.md | Removes the WeChat group join note and QR-code embed from the English README timeline section. |
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
prax211
referenced
this pull request
in cogniziocompany/LongHorizon-Harness
Sep 10, 2026
…te snippet
The reviewer this document argues for fired on this document and found a real
defect in it. Acting on all four findings.
1. SECURITY, and the reason for this commit. The illustrative "Job gate" snippet
read:
author_association == 'OWNER' || 'MEMBER' || 'COLLABORATOR'
In GitHub Actions expression syntax || short-circuits on the first truthy
operand and a non-empty string literal is truthy, so that condition is ALWAYS
TRUE. Anyone copying the snippet instead of the real file would have shipped a
reviewer with no write-access gate at all, invocable by any commenter -- while
the document's own verification section claimed outside contributors were
silently ignored. Replaced with the fully spelled comparison and a boxed
warning explaining why the condensed form is wrong, since it reads correctly
in English and that is exactly what makes it dangerous.
2. LongHorizon-Harness is no longer missing the workflow: PR #15, 7b27032, read
back from main at 2,491 bytes. Coverage table, "four repos" -> "three", and
rollout step 2 marked done with the evidence.
3. cognizioware-qa's default branch is master, not main. Noted in both the
coverage table and the rollout step, because targeting main there opens a PR
against a branch nobody merges.
4. cognizioware-hydra was framed as a coordination problem ("do not duplicate
#26"). That understated it: hydra-ci.yml has a deploy job on
[self-hosted, hydra-host] fired by push to main, targeting corsairai300, and
its own line 143 records that two such deploys wiped the CT110 node seeding.
CT110 runs every harness job and is on that host. It needs a zero-active-runs
window, not a Slack check.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
prax211
referenced
this pull request
in cogniziocompany/LongHorizon-Harness
Sep 10, 2026
* docs(handoff): reviewer coverage for the four repos that have none
Verified against the live GitHub API today rather than inherited from notes:
mcp-tools and mcp-cognizioware have claude-code.yml; powerplatform, hydra,
LongHorizon-Harness and qa return 404.
The finding that makes this small: CLAUDE_CODE_OAUTH_TOKEN already exists as an
ORG secret with visibility=all, so every repo can already read it. Nothing needs
provisioning per repo. The only missing artifact is a 65-line workflow file.
Records the four things that already bit us, so the next person does not
rediscover them:
- it does NOT fire on pull_request. Opening a PR reviews nothing; a human with
write access must comment @claude. People see no review and conclude the
workflow is broken when nobody tagged it.
- fetch-depth: 0 is load-bearing. A depth-1 clone has only the PR head, so the
reviewer cannot see main. On 2026-09-09 that made it call a real commit
non-existent and accuse correct work of being fabricated.
- timeout_minutes is not a valid action input and never applied. The real cap
is timeout-minutes at the job level, which matters because
cancel-in-progress: false lets a hung job block every later @claude on that
PR until the runner's 6-hour default.
- every invocation bills a personal subscription, not a metered key. Adding the
workflow costs nothing; each @claude does.
Rollout order is by blast radius: qa, then LongHorizon-Harness, then
powerplatform (where the customer-facing session-flow fixes ship), and hydra last
because PR #26 already adds it there and is gated behind BMAD by direction —
duplicating it would create a conflicting second workflow.
Also states how to verify, because a workflow file existing proves nothing: the
comment must be on a pull request, must literally contain @claude, and must come
from OWNER/MEMBER/COLLABORATOR — an outside contributor's comment is silently
ignored, and that silence looks exactly like a broken workflow.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs(handoff): act on the reviewer's findings — fix an always-true gate snippet
The reviewer this document argues for fired on this document and found a real
defect in it. Acting on all four findings.
1. SECURITY, and the reason for this commit. The illustrative "Job gate" snippet
read:
author_association == 'OWNER' || 'MEMBER' || 'COLLABORATOR'
In GitHub Actions expression syntax || short-circuits on the first truthy
operand and a non-empty string literal is truthy, so that condition is ALWAYS
TRUE. Anyone copying the snippet instead of the real file would have shipped a
reviewer with no write-access gate at all, invocable by any commenter -- while
the document's own verification section claimed outside contributors were
silently ignored. Replaced with the fully spelled comparison and a boxed
warning explaining why the condensed form is wrong, since it reads correctly
in English and that is exactly what makes it dangerous.
2. LongHorizon-Harness is no longer missing the workflow: PR #15, 7b27032, read
back from main at 2,491 bytes. Coverage table, "four repos" -> "three", and
rollout step 2 marked done with the evidence.
3. cognizioware-qa's default branch is master, not main. Noted in both the
coverage table and the rollout step, because targeting main there opens a PR
against a branch nobody merges.
4. cognizioware-hydra was framed as a coordination problem ("do not duplicate
#26"). That understated it: hydra-ci.yml has a deploy job on
[self-hosted, hydra-host] fired by push to main, targeting corsairai300, and
its own line 143 records that two such deploys wiped the CT110 node seeding.
CT110 runs every harness job and is on that host. It needs a zero-active-runs
window, not a Slack check.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
TON14
pushed a commit
to TON14/LongHorizon-Harness
that referenced
this pull request
Sep 17, 2026
update the readme file
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.