[lockfile-stats] Lockfile Statistics Report — 2026-09-04 (299 workflows) #58602
Closed
Replies: 1 comment
|
This discussion has been marked as outdated by Lockfile Statistics Analysis Agent. A newer discussion is available at Discussion #58866. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Overview
Analyzed 299 compiled
.github/workflows/*.lock.ymlfiles on 2026-09-04 (0 skipped/malformed). Total size 43.68 MB (45,801,415 bytes), avg 149.6 KB, min 90.0 KB, max 239.5 KB.Methodology: single-script compact JSON analysis (analyzer cached at
/tmp/gh-aw/cache-memory/scripts/lockfile_stats_v4.py, re-used across runs; full summary at/tmp/gh-aw/agent/lockfile-stats-summary.json).Key metrics
create_discussionworkflows w/ category detectedTrigger analysis
Top trigger combinations:
schedule+workflow_dispatch(209),workflow_dispatchonly (38),pull_request+schedule+workflow_dispatch(29),pull_request+workflow_dispatch(9).Top cron cadence:
0 0 */2 * *(every 2 days) dominates at 42 workflows; remaining schedules are mostly unique daily/weekly crons.Safe outputs analysis
Nearly universal safe-output types (present on ~293/299 workflows — the standard scaffold):
missing_data,missing_tool,noop,report_incomplete,create_report_incomplete_issue.Most common feature safe-outputs:
Discussion categories (92 workflows, 100% resolved): audits (79), announcements (5), artifacts (2), dev (2), research (2), general (1), daily-news (1).
New/rare integrations spotted this run:
jira_create_issueandlinear_create_issue(1 workflow each — new since 2026-09-03).Structural characteristics
Permission patterns (agent job)
All 299 workflows keep the agent job read-only on contents; write access is granted only via the union of other jobs. Union-of-all-jobs shows 299/299 workflows grant at least one write scope somewhere (mostly
issues: writeon all 299, andcontents: writeon 205) — consistent with the safe-outputs pattern where a separate privileged job performs writes on the agent's behalf.Engine distribution
Top models:
openai/gpt-5.4(48),copilot/mai-code-1-flash-picker(44),copilot/claude-sonnet-4.5(11),claude-sonnet-5(6).Tool & MCP patterns
MCP servers:
safeoutputs(299, universal),github(121),agenticworkflows(44),serena(25),mcpscripts(12),tavily(5),sentry/datadog/deepwiki/microsoftdocs/grafana/ast-grep(2-3 each), several single-use integrations (mempalace, graft, agentdb, ruflo).Most-used GitHub MCP read tools:
get_commit,get_file_contents,get_latest_release,list_commits,search_code(119 workflows each); PR-focused tool bundle (get_pull_request*,pull_request_read, etc.) at 113 workflows.Interesting findings
contents, with all writes routed through the safe-outputs mechanism — no direct write-capable agent jobs found.0 0 */2 * *alone accounts for 42 of ~91 distinct cron schedules, suggesting a shared default template rather than per-workflow tuning.create_discussionworkflows, 86%), meaning this very report's category is the norm, not the exception.jira_create_issueandlinear_create_issueeach showed up for the first time since 2026-09-03, alongside a net +1 lockfile and +1githubMCP server user.Historical trends (vs 2026-09-03)
Day-over-day change is minimal: one new workflow added (Copilot-engine, GitHub MCP, Jira/Linear issue creation), no regressions in engine/permission/safe-output detection.
Recommendations
audits(86%), confirm this is intentional taxonomy rather than a default nobody overrides.Warning
Firewall blocked 1 domain
The following domain was blocked by the firewall during workflow execution:
api.anthropic.comTo allow these domains, add them to the
network.allowedlist in your workflow frontmatter:See Network Configuration for more information.
All reactions