@htekdev/actions-debugger 1.0.78 → 1.0.79

package/errors/caching-artifacts/merge-queue-branch-cache-isolation.yml

@@ -0,0 +1,80 @@
+id: caching-artifacts-045
+title: "Merge queue (merge_group) runs cannot restore cache saved by pull_request runs"
+category: caching-artifacts
+severity: silent-failure
+tags:
+  - merge-queue
+  - merge_group
+  - cache-isolation
+  - branch-scoping
+  - cache-miss
+patterns:
+  - regex: 'Cache not found for input keys'
+    flags: i
+  - regex: 'gh-readonly-queue'
+    flags: i
+error_messages:
+  - "Cache not found for input keys:"
+  - "No cache found for key: Linux-node-"
+root_cause: |
+  GitHub Actions caches are branch-scoped. When the merge_group trigger fires,
+  GitHub creates a temporary branch named
+  gh-readonly-queue/main/pr-<number>-<sha> with a fresh merge commit. This
+  temporary branch has no prior cache entries, so every cache lookup is a cache
+  miss — even if the identical hashFiles() key was successfully saved during the
+  preceding pull_request run on the feature branch.
+
+  Branch-scoped cache fallback rules allow access to the base branch (e.g., main)
+  but NOT to arbitrary feature branches. The gh-readonly-queue/* branch is treated
+  as a new branch with no cache history and no access to feature-branch caches.
+  Only caches saved directly to main (or the configured merge target) are accessible
+  in merge queue runs.
+
+  The result: CI times effectively double. The pull_request run warms the cache
+  but the merge queue run cannot use it and must rebuild from scratch.
+fix: |
+  Seed caches on the base branch via push-triggered or scheduled workflows so
+  merge queue runs can restore from main's cache. Use restore-keys with a
+  branch-agnostic prefix to maximize fallback hit rate.
+fix_code:
+  - language: yaml
+    label: "Use restore-keys for base branch fallback"
+    code: |
+      - name: Cache dependencies
+        uses: actions/cache@v4
+        with:
+          path: ~/.npm
+          # Exact key includes lock file hash — works for push/PR
+          key: ${{ runner.os }}-node-${{ hashFiles('**/package-lock.json') }}
+          # restore-keys fall back to base branch cache (accessible in merge_group)
+          restore-keys: |
+            ${{ runner.os }}-node-
+
+  - language: yaml
+    label: "Seed base branch cache via push workflow"
+    code: |
+      # Separate workflow to warm cache on main after every merge
+      on:
+        push:
+          branches: [main]
+
+      jobs:
+        warm-cache:
+          runs-on: ubuntu-latest
+          steps:
+            - uses: actions/checkout@v4
+            - uses: actions/setup-node@v4
+              with:
+                node-version: '20'
+                cache: npm
+            - run: npm ci
+prevention:
+  - "Seed caches on your base branch (main) so merge queue runs can restore from it."
+  - "Use branch-agnostic restore-keys prefixes — they survive the gh-readonly-queue/* branch context."
+  - "Avoid cache keys that include branch name or github.head_ref — they will never match in merge queue."
+  - "setup-node/setup-python cache: integration automatically includes restore-keys and handles this gracefully."
+docs:
+  - url: "https://docs.github.com/en/actions/writing-workflows/choosing-when-your-workflow-runs/events-that-trigger-workflows#merge_group"
+    label: "GitHub Actions: merge_group trigger"
+  - url: "https://docs.github.com/en/actions/writing-workflows/choosing-what-your-workflow-does/caching-dependencies-to-speed-up-workflows#restrictions-for-accessing-a-cache"
+    label: "GitHub Actions: Cache branch access restrictions"

package/errors/concurrency-timing/rerun-cancel-original-missing-run-attempt.yml

@@ -0,0 +1,67 @@
+id: concurrency-timing-039
+title: "Manual re-run cancels original run when concurrency group omits github.run_attempt"
+category: concurrency-timing
+severity: error
+tags:
+  - concurrency
+  - re-run
+  - cancel-in-progress
+  - run-attempt
+  - manual-rerun
+patterns:
+  - regex: 'This run was cancelled'
+    flags: i
+  - regex: 'run_attempt'
+    flags: i
+error_messages:
+  - "This run was cancelled."
+  - "A run for this workflow is already in progress."
+root_cause: |
+  When cancel-in-progress: true is set and the concurrency group key does not
+  include github.run_attempt, every manual re-run of a workflow is treated as a
+  duplicate of the original run (since run_id and ref are the same). The re-run
+  immediately cancels the still-executing original attempt, or the concurrency
+  group queues the re-run and cancels any pending run — which may be the active
+  original.
+
+  The run_id is stable across re-runs but run_attempt increments (1, 2, 3...).
+  Without run_attempt in the concurrency group key, GitHub has no way to
+  distinguish re-run attempt 2 from the still-running attempt 1.
+
+  This produces confusing CI failures: a developer clicks "Re-run jobs" on a
+  failed workflow and immediately sees the original (or the re-run itself)
+  cancelled by the concurrency system.
+fix: |
+  Include github.run_attempt in the concurrency group key so each attempt is
+  treated as a distinct slot. This allows re-runs to coexist with or cleanly
+  supersede the original without unexpected cancellations.
+fix_code:
+  - language: yaml
+    label: "Include run_attempt in concurrency group key"
+    code: |
+      # WRONG: re-runs and originals share the same concurrency slot
+      # concurrency:
+      #   group: ${{ github.workflow }}-${{ github.ref }}
+      #   cancel-in-progress: true
+
+      # CORRECT: each attempt gets its own slot, preventing cross-attempt cancellation
+      concurrency:
+        group: ${{ github.workflow }}-${{ github.ref }}-${{ github.run_attempt }}
+        cancel-in-progress: true
+
+  - language: yaml
+    label: "Alternative: cancel-in-progress only for push/PR, not re-runs"
+    code: |
+      concurrency:
+        group: ${{ github.workflow }}-${{ github.ref }}
+        # Only cancel-in-progress for first attempt; re-runs are never cancelled
+        cancel-in-progress: ${{ github.run_attempt == 1 }}
+prevention:
+  - "Always include github.run_attempt in concurrency group keys when cancel-in-progress: true is set."
+  - "Test re-run behavior explicitly — concurrency cancellation bugs only surface when you manually re-run."
+  - "Consider using cancel-in-progress: ${{ github.run_attempt == 1 }} to preserve re-run integrity."
+docs:
+  - url: "https://docs.github.com/en/actions/writing-workflows/workflow-syntax-for-github-actions#concurrency"
+    label: "GitHub Actions: Concurrency syntax"
+  - url: "https://docs.github.com/en/actions/writing-workflows/choosing-what-your-workflow-does/evaluate-expressions-in-workflows-and-actions#github-context"
+    label: "GitHub Actions: github.run_attempt context"

package/errors/silent-failures/fromjson-object-stored-in-env-stringifies.yml

@@ -0,0 +1,77 @@
+id: silent-failures-074
+title: "fromJSON() object or array stored in env: block silently becomes '[object Object]'"
+category: silent-failures
+severity: silent-failure
+tags:
+  - fromjson
+  - env-context
+  - type-coercion
+  - json-parsing
+  - stringification
+patterns:
+  - regex: 'env\s*:.*\$\{\{\s*fromJSON\('
+    flags: ims
+error_messages:
+  - "[object Object]"
+  - "Unexpected token 'o', \"[object O\"... is not valid JSON"
+root_cause: |
+  The fromJSON() function in GitHub Actions expressions parses a JSON string
+  into a typed value — object, array, boolean, or number. However, when the
+  result is assigned to an env: variable, the Actions runner coerces the value
+  to a string using JavaScript's default toString(). For objects this produces
+  the literal string [object Object]; for arrays it produces a comma-joined
+  string (e.g., a,b,c).
+
+  The step exits with code 0 and the env variable appears set, but its value is
+  the stringified form rather than the JSON structure the developer expected.
+  Subsequent steps that reference ${{ env.CONFIG }} or echo the variable and
+  pipe it through jq will receive [object Object] and fail to parse — or silently
+  produce wrong results.
+
+  Booleans and numbers survive correctly: fromJSON('true') in env: produces the
+  string 'true', and fromJSON('42') produces '42' — these are the only safe uses.
+fix: |
+  Pass structured JSON data between steps via GITHUB_OUTPUT (as a raw JSON
+  string), then consume the stored string with fromJSON() in downstream
+  expression contexts. Never store objects or arrays via fromJSON() in env: blocks.
+fix_code:
+  - language: yaml
+    label: "Store JSON in GITHUB_OUTPUT, consume via fromJSON in expressions"
+    code: |
+      jobs:
+        build:
+          runs-on: ubuntu-latest
+          steps:
+            # WRONG: fromJSON object in env: becomes '[object Object]'
+            # - name: Bad config
+            #   env:
+            #     CONFIG: ${{ fromJSON(inputs.config_json) }}
+            #   run: echo "$CONFIG" | jq .host  # fails: '[object Object]' is not JSON
+
+            # CORRECT: store raw JSON string in GITHUB_OUTPUT
+            - name: Store config
+              id: cfg
+              run: echo 'config=${{ inputs.config_json }}' >> $GITHUB_OUTPUT
+
+            # Access fields with fromJSON() directly in expression context
+            - name: Use config
+              run: |
+                echo "Host: ${{ fromJSON(steps.cfg.outputs.config).host }}"
+                echo "Port: ${{ fromJSON(steps.cfg.outputs.config).port }}"
+
+            # Scalar fromJSON values are safe in env: (boolean, number)
+            - name: Scalar fromJSON is fine
+              env:
+                IS_PROD: ${{ fromJSON(inputs.is_prod) }}
+                TIMEOUT: ${{ fromJSON(inputs.timeout) }}
+              run: echo "prod=$IS_PROD timeout=$TIMEOUT"
+prevention:
+  - "Never assign fromJSON() results that produce objects or arrays to env: variables."
+  - "Use fromJSON() inline in expression contexts (${{ fromJSON(x).field }}) rather than materializing into env vars."
+  - "Store JSON blobs as raw strings in GITHUB_OUTPUT and parse with jq in shell or fromJSON() in expressions."
+  - "For object env vars, keep the value as a raw JSON string and parse in shell with jq or python -c."
+docs:
+  - url: "https://docs.github.com/en/actions/writing-workflows/choosing-what-your-workflow-does/evaluate-expressions-in-workflows-and-actions#fromjson"
+    label: "GitHub Actions: fromJSON expression function"
+  - url: "https://docs.github.com/en/actions/writing-workflows/choosing-what-your-workflow-does/passing-information-between-jobs"
+    label: "GitHub Actions: Passing information between jobs"

package/errors/triggers/event-types-unknown-value-silently-ignored.yml

@@ -0,0 +1,89 @@
+id: triggers-053
+title: "Unknown or misspelled values in on.<event>.types[] are silently ignored — workflow never triggers"
+category: triggers
+severity: silent-failure
+tags:
+  - types-filter
+  - trigger
+  - typo
+  - pull-request
+  - silent-failure
+patterns:
+  - regex: 'types\s*:\s*\[.*(?:syncronize|synchronised|synchronize d|re-opened|labeld|assign)\]'
+    flags: i
+  - regex: 'types\s*:\s*\n(?:\s+-\s+\w+\n)*\s+-\s+(?!opened|closed|reopened|synchronize|labeled|unlabeled|assigned|unassigned|review_requested|ready_for_review|converted_to_draft|locked|unlocked|milestoned|demilestoned|edited)\w'
+    flags: im
+error_messages:
+  - "Workflow is not triggered"
+  - "No runs found"
+root_cause: |
+  GitHub Actions silently discards unrecognized values in the types: filter
+  array for pull_request, pull_request_target, and other typed events. There is
+  no validation error, no schema warning, and no run log entry — the workflow
+  simply never fires for the intended activity type.
+
+  The most common cause is a typo in a well-known type name:
+    - syncronize  (correct: synchronize)
+    - re-opened   (correct: reopened)
+    - labeld      (correct: labeled)
+    - closed      (correct: closed — note: not 'merged')
+
+  A second common cause is listing only the new type while forgetting that the
+  default types are no longer active. For example, types: [synchronize] drops
+  the implicit opened and reopened behaviors, so a newly opened PR will not
+  trigger CI at all.
+
+  actionlint >=0.7.0 can detect unknown type values but is not always used in
+  local development workflows.
+fix: |
+  Verify all type values against the official list for each event. Include all
+  intended types explicitly — once types: is specified, the defaults no longer
+  apply.
+fix_code:
+  - language: yaml
+    label: "Correct types for pull_request CI — always include all intended types"
+    code: |
+      on:
+        pull_request:
+          # WRONG examples (silently never trigger):
+          # types: [opened, syncronize, re-opened]   # typos
+          # types: [synchronize]                     # drops opened/reopened
+
+          # CORRECT: include all intended activity types
+          types:
+            - opened
+            - synchronize
+            - reopened
+
+  - language: yaml
+    label: "All valid pull_request types for reference"
+    code: |
+      on:
+        pull_request:
+          types:
+            # Commonly needed for CI:
+            - opened          # new PR created
+            - synchronize     # new commit pushed to PR branch
+            - reopened        # closed PR re-opened
+            # Useful for label-gated workflows:
+            - labeled
+            - unlabeled
+            # Draft/review flow:
+            - ready_for_review
+            - converted_to_draft
+            # Assignment:
+            - assigned
+            - unassigned
+            - review_requested
+prevention:
+  - "Run actionlint (>=0.7.0) in CI to detect unknown type values — it reports them as errors."
+  - "When adding types:, explicitly list ALL types you need — defaults no longer apply once types: is present."
+  - "After changing types:, trigger one manual test run per type to confirm the filter is working."
+  - "Use the GitHub Actions VS Code extension — it underlines unknown type values as warnings."
+docs:
+  - url: "https://docs.github.com/en/actions/writing-workflows/choosing-when-your-workflow-runs/events-that-trigger-workflows#pull_request"
+    label: "GitHub Actions: pull_request event types"
+  - url: "https://docs.github.com/en/actions/writing-workflows/choosing-when-your-workflow-runs/events-that-trigger-workflows#pull_request_target"
+    label: "GitHub Actions: pull_request_target event types"
+  - url: "https://rhysd.github.io/actionlint/"
+    label: "actionlint: GitHub Actions linter"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@htekdev/actions-debugger",
-  "version": "1.0.78",
+  "version": "1.0.79",
   "description": "65+ real GitHub Actions errors, queryable by agents. CLI + MCP server + Copilot skills + error database.",
   "type": "module",
   "main": "./dist/index.js",