feat(token-optimizer): migrate prompt to gh-proxy mode with bash+jq efficiency guidance

[Copilot]

The token optimizer workflow prompt still referenced vague "cli-proxy" guidance and lacked concrete instructions for token-efficient GitHub API access, despite the frontmatter already declaring mode: gh-proxy.

Changes

• New "Data Access Guidelines" section — explicitly declares no GitHub MCP tools are available; provides ✅/❌ examples showing gh api --jq filtering vs. loading full unfiltered payloads into context:

  # ✅ Extract only needed fields
  gh api "repos/$REPO/actions/workflows/my-workflow.yml/runs?per_page=10" \
    --jq '.workflow_runs[] | {id, name, conclusion, run_started_at}'
  
  # ❌ Drops entire response into context
  gh api "repos/$REPO/actions/workflows/my-workflow.yml/runs"

• Phase 2 analysis matrix — "Tool usage" row now references gh api --jq '.content' | base64 -d | sed -n … to extract only frontmatter instead of the generic "via cli-proxy" phrase.

• Phase 3 (Read Workflow Source) — replaced the paragraph with concrete bash snippets using gh api --jq + awk one-liners to slice out frontmatter or prompt body independently, avoiding full-file loads.

[Copilot]

Pull request overview

Updates the Copilot Token Optimizer workflow prompt to better match mode: gh-proxy usage and to encourage token-efficient GitHub API access patterns (bash + gh api --jq + targeted slicing).

Changes:

• Add a new “Data Access Guidelines” section with gh api --jq examples intended to reduce token waste from large JSON payloads.
• Update Phase 2/3 instructions to reference extracting only frontmatter/prompt slices from workflow sources instead of generic “cli-proxy” wording.
• Pass GH_AW_GITHUB_REPOSITORY into the “Interpolate variables and render templates” step in the locked workflow.

Show a summary per file

File	Description
.github/workflows/copilot-token-optimizer.md	Adds/updates prompt guidance for token-efficient gh api --jq usage and targeted extraction of workflow source sections.
.github/workflows/copilot-token-optimizer.lock.yml	Adds GH_AW_GITHUB_REPOSITORY env var to the interpolate/render step environment.

Copilot's findings

Tip

Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comments suppressed due to low confidence (1)

.github/workflows/copilot-token-optimizer.md:226

• This awk command does not extract only the prompt body after the closing ---. As written, it starts printing after the first ---, so it includes the entire frontmatter (and the closing delimiter) in the output. Update the command to begin output only after the second --- delimiter so the “prompt body only” instruction is accurate.

# Extract the prompt body only (everything after the closing ---)
gh api "repos/$REPO/contents/$WF_PATH" --jq '.content' | base64 -d \
  | awk 'f; /^---$/{f=1}'

• Files reviewed: 2/2 changed files
• Comments generated: 4