fix(safe-outputs): re-derive patch/bundle paths and stop double-writing infrastructure fields (#37299)

* fix patch/bindle file paths

* remove all mention of patch_path and bundle_path

* recompile workflow

* recompile workflow

* Update wasm golden files for GH_AW_TIMEOUT_MINUTES env var

PR #37262 (e76a712) added GH_AW_TIMEOUT_MINUTES to the agent step env
in copilot_engine_execution.go but did not refresh the wasm golden
fixtures. Regenerated via 'make update-wasm-golden'.

* docs: replace Effective Tokens terminology with AI Credits (AIC)

* docs: rename GH_AW_DEFAULT_MAX_EFFECTIVE_TOKENS → GH_AW_DEFAULT_MAX_AI_CREDITS

* clean old stuff

Author: dsyme

.changeset/patch-bump-awf-v0-25-44.md

@@ -29,7 +29,7 @@ description: Agentic workflow specific frontmatter fields for GitHub Agentic Wor
   - Top-level field mapped to `apiProxy.maxRuns`
   - Supported by all engines
 - **`max-turns:`** - AWF turn cap applied consistently across all agentic engines (integer or expression, e.g. `${{ inputs.max-turns }}`). The engine-level `engine.max-turns` is a deprecated alias kept for backward compatibility — prefer this top-level field. Not supported by the `gemini` engine.
-- **`max-effective-tokens:`** - Per-run effective-token (ET) budget enforced by the AWF firewall (integer or `K`/`M` short-form string like `100M`; default `25000000`). Set a negative value to disable enforcement and token steering. See [token-optimization.md](token-optimization.md).
+- **`max-ai-credits:`** - Per-run effective-token (ET) budget enforced by the AWF firewall (integer or `K`/`M` short-form string like `100M`; default `25000000`). Set a negative value to disable enforcement and token steering. See [token-optimization.md](token-optimization.md).
 - **`max-daily-ai-credits:`** - Per-user 24-hour ET guardrail: activation blocks execution once the triggering user's aggregated ET for this workflow over the last 24h exceeds the threshold (integer or `K`/`M` short-form string, or `-1`). Enabled by default with a system default threshold; set `-1` to disable or an explicit value to override. See [token-optimization.md](token-optimization.md).
 - **`user-rate-limit:`** - Rate limiting configuration to prevent users from triggering the workflow too frequently (object)
   - **`max-runs-per-window:`** - Maximum runs allowed per user per time window (required, integer 1-10)

.github/aw/syntax-agentic.md

@@ -363,11 +363,11 @@ To maximize cache hits:
 
 Two top-level frontmatter fields enforce ET budgets directly, independent of the techniques above. Both accept an integer or a `K`/`M` short-form string (e.g. `100M`, `500K`).
 
-- **`max-effective-tokens:`** — Per-run ET budget enforced by the AWF firewall/API proxy (default `25000000`). The agent is steered to stay within budget; set a negative value to disable enforcement and steering.
+- **`max-ai-credits:`** — Per-run ET budget enforced by the AWF firewall/API proxy (default `25000000`). The agent is steered to stay within budget; set a negative value to disable enforcement and steering.
 - **`max-daily-ai-credits:`** — Per-user 24-hour guardrail. At activation, gh-aw sums the triggering user's ET across their runs of this workflow over the last 24 hours and blocks execution once the total exceeds the threshold. Enabled by default with a system default threshold; set `-1` to disable, or an explicit value to override the default.
 
 ```yaml
-max-effective-tokens: 100M        # per-run cap (short-form string)
+max-ai-credits: 100M        # per-run cap (short-form string)
 max-daily-ai-credits: 500M  # per-user 24h cap; -1 disables
 ```
 

.github/aw/token-optimization.md

@@ -40,7 +40,7 @@ safe-outputs:
 timeout-minutes: 30
 # Raised above 25M default: meta-orchestrator does deep analysis across all agents.
 # Caveman experiment aims to reduce by ≥20%; set ceiling to catch runaway loops.
-max-effective-tokens: 40000000
+max-ai-credits: 40000000
 experiments:
   prompt_compression:
     variants: [verbose, caveman]