Fix gh-aw shell-injection warnings and update generated locks

[Evangelink]

What

Fix gh aw compile --strict shell-injection warnings on 3 agentic workflows and update generated lock files.

Source workflow changes (3 files)

• .github/workflows/build-failure-analysis.md
• .github/workflows/build-failure-analysis-command.md
• .github/workflows/issue-arborist.md

Each one moves ${{ ... }} expressions out of run: scripts into step-level env: blocks, satisfying the strict-mode shell-injection check. The expressions covered are: steps.find-binlog.outputs.path, github.repository, inputs.pr-number, github.event.issue.number, steps.build.outcome, github.workspace, and the resolved PR head SHA fallback chain.

Generated changes (regenerated outputs)

• .github/aw/actions-lock.json — gh-aw-actions/setup v0.72.1 → v0.74.8; new firewall image SHAs (0.25.41, 0.25.46, 0.25.49)
• .github/workflows/agentics-maintenance.yml and ~18 other .lock.yml files — pure regeneration from gh aw compile after the runtime version bump; no source .md edits.
• .github/workflows/msbuild-quality-review.lock.yml — also regenerated; supersedes the partial regen from Regenerate stale msbuild-quality-review.lock.yml #8455 (now consistently at v0.74.8).

Remaining gh aw compile --strict output

Only one informational warning is left and is intentionally deferred to a follow-up PR:

Detected 6 slash_command entries; 6 are not using centralized routing. Consider setting on.slash_command.strategy: centralized to reduce duplicate triggers and route through agentic_commands.yml.

How to verify

gh aw compile --strict

should emit 0 error(s) and only the slash-command centralization suggestion above.

[Copilot]

Pull request overview

Updates agentic workflow sources and regenerates gh-aw lock artifacts to eliminate gh aw compile --strict shell-injection warnings and to align all generated workflows/locks with gh-aw v0.74.8.

Changes:

• Move ${{ ... }} GitHub expressions out of run: scripts into step-level env: in the 3 affected agentic workflow sources.
• Regenerate .lock.yml workflows and update .github/aw/actions-lock.json to reflect gh-aw v0.74.8 and updated pinned action/container SHAs.
• Refresh generated agentics-maintenance.yml content (including the newly generated forecast maintenance operation).

Show a summary per file

File	Description
.github/workflows/build-failure-analysis.md	Moves step outputs/context expressions from run: into env: for strict-mode shell-injection compliance.
.github/workflows/build-failure-analysis-command.md	Same strict-mode fix as above for the slash-command variant workflow.
.github/workflows/issue-arborist.md	Moves github.repository expression out of the curl run: script into step env:.
.github/workflows/repo-historian.lock.yml	Regenerated lock workflow with updated gh-aw version, pinned actions/containers, and updated generated harness snippets.
.github/workflows/issue-arborist.lock.yml	Regenerated lock workflow with updated gh-aw version, pinned actions/containers, and generated env-expression handling.
.github/workflows/msbuild-quality-review.lock.yml	Regenerated lock workflow for gh-aw v0.74.8; includes updated pinned dependencies and regenerated job steps.
.github/workflows/daily-malicious-code-scan.lock.yml	Regenerated lock workflow for gh-aw v0.74.8; updates pinned actions/containers (incl. CodeQL upload-sarif).
.github/workflows/agentics-maintenance.yml	Regenerated maintenance workflow for gh-aw v0.74.8; includes the new generated forecast operation path.
.github/aw/actions-lock.json	Updates pinned action/container entries to match regenerated workflows (gh-aw-actions/setup v0.74.8, updated SHAs/images).

Copilot's findings

• Files reviewed: 19/26 changed files
• Comments generated: 0

[Evangelink]

✅ 21/21 dimensions clean — no findings.

This PR is a textbook-correct security fix for shell injection vulnerabilities in GitHub Actions workflows. The changes follow the industry-standard pattern of moving GitHub Actions expressions (${{ ... }}) from run: script blocks into step-level env: blocks, preventing potential injection attacks through repository names or issue numbers.

What was fixed:

• 3 workflow source files (.md) now use environment variables instead of inline expressions
• 23 generated lock files updated after gh-aw runtime bump (v0.72.1 → v0.74.8)

Security impact: The vulnerable pattern allowed expressions like ${{ github.repository }} to be interpolated directly into shell commands. If an attacker could control the repository name (e.g., via a fork or malicious issue), they could potentially inject shell metacharacters. The new pattern evaluates expressions in the GitHub Actions runtime before the shell sees them, eliminating this attack vector.

Code quality: The transformation is applied consistently across all three affected workflows with proper variable naming (GH_AW_* prefix) and complete coverage of all vulnerable expressions.

The generated lock file changes are mechanical and correct — they reflect the runtime version bump and updated firewall image SHAs.

Post-merge note: This PR is already merged. This review confirms the changes are correct and complete.

Generated by Expert Code Review (on open) for issue #8458 · ● 3.8M