This change stops agentic workflows from opening GitHub issues when a
run fails due to workflow/tooling conditions such as missing-tool
reports or token-budget exhaustion. It applies the repository-wide
suppression suggested by the failure issue itself, so these runs can
still fail without creating issue noise.
- **What changed**
- Added `safe-outputs.report-failure-as-issue: false` to each top-level
agentic workflow source under `.github/workflows/*.md`
- Regenerated the corresponding compiled `.lock.yml` workflows so the
runtime configuration matches the source frontmatter
- **Effect**
- Agentic workflow runs continue to report failure in Actions
- Automatic `[aw] ... failed` issue creation is disabled for these
workflows
- Existing safe outputs such as `noop` and `missing-tool` remain
unchanged
- **Scope**
- Applied consistently across the repository’s top-level agentic
workflows, including `zipt-code-reviewer`, `build-warning-fixer`,
`code-conventions-analyzer`, `workflow-suggestion-agent`, and related
workflows
- **Configuration pattern**
```yaml
safe-outputs:
report-failure-as-issue: false
create-issue:
...
missing-tool:
create-issue: true
noop:
report-as-issue: false
```
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Monthly agentic workflow that discovers community-contributed SMT-LIB
benchmark repositories on GitHub, excluding official distributions from
smtlib.org and Zenodo, and posts a categorised summary as a GitHub
Discussion.
## Workflow steps
- **Exclusion list** — scrapes `smtlib.cs.uiowa.edu` + paginates the
[Zenodo SMT-LIB community](https://zenodo.org/communities/smt-lib/) (up
to 300 records) to extract official repo URLs; hard-excludes `SMT-LIB/*`
and `SMT-Competition/*` orgs
- **GitHub search** — 7 search strategies (topic tags
`smtlib`/`smt-lib`/`smt2`, filename/path patterns, README mentions)
scoped to repos pushed since the last run (90-day window on first run)
- **Filter + classify** — removes official sets and noise (homework
repos, empty repos); categorises survivors into: solver evaluation,
verification, security/CTF, theory/logic, tool output, education, other
- **Discussion report** — posts `[SMT-LIB Benchmarks] Community
Benchmark Repository Survey — [Month YYYY]`; older discussions
auto-closed after 90 days
- **Cache** — stores classified repo sets and run date so subsequent
runs are incremental
## Files
- `.github/workflows/smtlib-benchmark-finder.md` — workflow definition
- `.github/workflows/smtlib-benchmark-finder.lock.yml` — compiled GitHub
Actions YAML (`gh aw compile`)
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
