You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Over the last 7 days, 97 firewall-enabled workflow runs across github/gh-aw were analyzed, covering 5,236 network requests. The overall block rate was low at 1.22% (64 blocked requests), indicating that firewall allowlists are largely well-tuned to workflow needs. The most notable blocked traffic came from registry.npmjs.org (17 blocks, mostly in PureLock, Dead Code Removal Agent, and the Daily AIC Consumption Report), and 44 requests to unidentified/redacted domains, concentrated in PureLock (41 blocked) and Dead Code Removal Agent (11 blocked).
No DIFC integrity-filtered events were found in the last 7 days — the DIFC gateway data source returned zero runs, and the cache-memory snapshot confirms an empty history as of this run. This suggests either no integrity/secrecy violations were triggered, or DIFC filtering is not yet actively engaged for the monitored workflows.
Cross-cutting theme: PureLock and Dead Code Removal Agent are the two workflows generating the most firewall blocks, both tied to package-registry access (npmjs.org) — likely legitimate dependency-resolution traffic that isn't yet allowlisted. No workflow appeared in both firewall blocks and DIFC filtering since DIFC data was empty this cycle.
🔥 Firewall Analysis
Key Firewall Metrics
Metric
Value
Workflows analyzed (firewall-enabled)
97
Total network requests monitored
5,236
✅ Allowed requests
5,172
🚫 Blocked requests
64
Block rate
1.22%
Total unique blocked domains
4
📈 Firewall Request Trends
All analyzed runs fall within a single calendar day (2026-08-20) within the 7-day window, so the trend chart shows one data point: 5,172 allowed vs. 64 blocked requests. No multi-day trend is available yet since the collected sample only spans the current day's runs; future runs will build out the historical trend as more days accumulate in cache-memory.
Top Blocked Domains
The largest block category is an unidentified/redacted domain bucket (44 blocks) spread across 8 workflows, followed by registry.npmjs.org:443 (17 blocks) tied to package installs in code-analysis workflows. storage.googleapis.com and proxy.golang.org each saw minimal, isolated blocks tied to single workflows.
No individual rule-hit counts were reported in the per-run policy_analysis payloads (rule_hits arrays were empty across all 97 audited runs), so rule-level attribution beyond the policy summary is not available this cycle.
Investigate the "unknown/redacted" blocked category (44 hits, 8 workflows) — these are the largest source of blocks but lack domain attribution in the audit output; pull raw firewall logs for PureLock and Dead Code Removal Agent runs to identify the actual target domains and confirm they aren't sensitive exfiltration attempts.
Consider allowlisting registry.npmjs.org:443 for PureLock, Dead Code Removal Agent, and the Daily AIC Consumption Report workflows if npm package installs are an expected part of their operation — 17 blocks suggest legitimate but currently-restricted dependency resolution.
PureLock's high block count (41/208 requests, ~20%) warrants a dedicated review of its network policy; it is the single largest source of blocked traffic and may need targeted allowlist tuning rather than broad relaxation.
Low-frequency singleton blocks (Architecture Guardian, CLI Consistency Checker, Daily Credit Limit Test, Daily Malicious Code Scan Agent — 1 block each) are likely one-off noise; monitor but no immediate action needed unless the pattern recurs.
🔒 DIFC Integrity Analysis
No DIFC integrity-filtered events were found in the last 7 days. The /tmp/gh-aw/agent/integrity/filtered-logs.json source returned zero matching runs, and the warm-start cache snapshot (updated_at: 2026-08-20T16:11:28Z) confirms an empty history, so DIFC chart generation (Phase 4) was skipped per the report procedure.
💡 DIFC Tuning Recommendations
Confirm that DIFC gateway logging is correctly enabled and reporting for all monitored workflows, since a fully empty result across 7 days may indicate either a clean bill of health or a data-collection gap — verify with a manual spot-check of one recent run's gateway logs.
Re-run this analysis on the next scheduled cycle to confirm whether the absence of DIFC events is a consistent trend or a one-off gap in the current sample window.
Generated by the Daily Security Observability workflow (consolidated from Daily Firewall Reporter + Daily DIFC Analyzer) Analysis window: Last 7 days | Repository: github/gh-aw Run: https://github.com/github/gh-aw/actions/runs/32389542144
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
Executive Summary
Over the last 7 days, 97 firewall-enabled workflow runs across github/gh-aw were analyzed, covering 5,236 network requests. The overall block rate was low at 1.22% (64 blocked requests), indicating that firewall allowlists are largely well-tuned to workflow needs. The most notable blocked traffic came from
registry.npmjs.org(17 blocks, mostly in PureLock, Dead Code Removal Agent, and the Daily AIC Consumption Report), and 44 requests to unidentified/redacted domains, concentrated in PureLock (41 blocked) and Dead Code Removal Agent (11 blocked).No DIFC integrity-filtered events were found in the last 7 days — the DIFC gateway data source returned zero runs, and the cache-memory snapshot confirms an empty history as of this run. This suggests either no integrity/secrecy violations were triggered, or DIFC filtering is not yet actively engaged for the monitored workflows.
Cross-cutting theme: PureLock and Dead Code Removal Agent are the two workflows generating the most firewall blocks, both tied to package-registry access (
npmjs.org) — likely legitimate dependency-resolution traffic that isn't yet allowlisted. No workflow appeared in both firewall blocks and DIFC filtering since DIFC data was empty this cycle.🔥 Firewall Analysis
Key Firewall Metrics
📈 Firewall Request Trends
All analyzed runs fall within a single calendar day (2026-08-20) within the 7-day window, so the trend chart shows one data point: 5,172 allowed vs. 64 blocked requests. No multi-day trend is available yet since the collected sample only spans the current day's runs; future runs will build out the historical trend as more days accumulate in cache-memory.
Top Blocked Domains
The largest block category is an unidentified/redacted domain bucket (44 blocks) spread across 8 workflows, followed by
registry.npmjs.org:443(17 blocks) tied to package installs in code-analysis workflows.storage.googleapis.comandproxy.golang.orgeach saw minimal, isolated blocks tied to single workflows.Most Frequently Blocked Domains
Policy Rule Attribution
📋 Policy: 11 rules, SSL Bump disabled, DLP disabled
No individual rule-hit counts were reported in the per-run policy_analysis payloads (rule_hits arrays were empty across all 97 audited runs), so rule-level attribution beyond the policy summary is not available this cycle.
View Detailed Request Patterns by Workflow
View Complete Blocked Domains List
🔒 Firewall Security Recommendations
registry.npmjs.org:443for PureLock, Dead Code Removal Agent, and the Daily AIC Consumption Report workflows if npm package installs are an expected part of their operation — 17 blocks suggest legitimate but currently-restricted dependency resolution.🔒 DIFC Integrity Analysis
No DIFC integrity-filtered events were found in the last 7 days. The
/tmp/gh-aw/agent/integrity/filtered-logs.jsonsource returned zero matching runs, and the warm-start cache snapshot (updated_at: 2026-08-20T16:11:28Z) confirms an empty history, so DIFC chart generation (Phase 4) was skipped per the report procedure.💡 DIFC Tuning Recommendations
Generated by the Daily Security Observability workflow (consolidated from Daily Firewall Reporter + Daily DIFC Analyzer)
Analysis window: Last 7 days | Repository: github/gh-aw
Run: https://github.com/github/gh-aw/actions/runs/32389542144
All reactions