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
Stacked on PR #19 (decouple Gas City harness). Three additions:
tags field on IR PlanOperation — optional readonly string[] for prioritization, filtering, categorization
Critical-check prioritization in runtime topoSort — critical-tagged ops scheduled first among ready siblings (stable sort, input order preserved within groups)
gc-watchdog skill — project-local skill at .agents/skills/gc-watchdog/ with standalone script and exit condition
Changes
File
Change
packages/ir/src/plan.ts
Add tags?: readonly string[] to PlanOperation
packages/sdk/src/convert.ts
Pass spec.tags through in convertOperation
packages/runtime/src/internal/topo.ts
sortReadyByPriority() — critical first among ready siblings
Adds tags to IR operations, prioritizes "critical" checks among ready siblings, and adds a gc-watchdog skill/script to exit cleanly when Gas City is idle. Also brings in CLI, checks, compilers, and docs; addresses Linear sv-bni by delivering the deferred critical-check prioritization from ADR-007.
New Features
IR/SDK/Runtime: PlanOperation.tags?: readonly string[]; SDK propagates spec.tags; topoSort schedules "critical" first (stable within groups).
Agents: .agents/skills/gc-watchdog/ with watchdog.sh (filters wisp/nudge; exits only on clean idle with healthy mayor).
Docs/Tooling: ADR-007/008 and architecture/user docs (waves 03–15); ESLint 9 flat config and Codacy config updates.
Written for commit 238bef0. Summary will update on new commits.
CodeAnt-AI Description
Add operation tags, prioritize critical work, and provide a Gas City watchdog
What Changed
Operations can now carry optional tags such as critical and security, including empty tag lists.
Critical-tagged operations run before non-critical operations when ready, while preserving order within each priority group.
Added a watchdog skill and standalone script that reports Gas City health and exits only when real work is complete and the city is healthy.
Added documentation describing the watchdog checks, reporting, and clean-idle exit conditions.
Impact
✅ Critical checks run sooner ✅ Tags remain available for operation filtering and categorization ✅ Clearer visibility into Gas City health and remaining work
Every time you make a pull request, our system automatically looks through it. We check for security issues, mistakes in how you're setting up your infrastructure, and common code problems. We do this to make sure your changes are solid and won't cause any trouble later.
Talking to CodeAnt AI
Got a question or need a hand with something in your pull request? You can easily get in touch with CodeAnt AI right here. Just type the following in a comment on your pull request, and replace "Your question here" with whatever you want to ask:
@codeant-ai ask: Your question here
This lets you have a chat with CodeAnt AI about your pull request, making it easier to understand and improve your code.
Example
@codeant-ai ask: Can you suggest a safer alternative to storing this secret?
Preserve Org Learnings with CodeAnt
You can record team preferences so CodeAnt AI applies them in future reviews. Reply directly to the specific CodeAnt AI suggestion (in the same thread) and replace "Your feedback here" with your input:
@codeant-ai: Your feedback here
This helps CodeAnt AI learn and adapt to your team's coding style and standards.
Example
@codeant-ai: Do not flag unused imports.
Retrigger review
Ask CodeAnt AI to review the PR again, by typing:
@codeant-ai: review
Check Your Repository Health
To analyze the health of your code repository, visit our dashboard at https://app.codeant.ai. This tool helps you identify potential issues and areas for improvement in your codebase, ensuring your repository maintains high standards of code health.
@ThePlenkov, you've reached your PR review limit, so we couldn't start this review.
Next review available in:47 minutes
Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).
How can I continue?
After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.
To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.
How do review limits work?
CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.
For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.
The reason will be displayed to describe this comment to others. Learn more.
Review Summary
This PR adds a tags field for operation prioritization and a gc-watchdog monitoring skill. The implementation is well-structured with comprehensive test coverage (3 IR validation tests, 2 SDK conversion tests, 2 runtime topo tests all passing).
Critical Issue
Shell script logic error: The regex pattern in watchdog.sh line 22 expects exactly 4 characters after sv- but attempts to filter sv-wisp-XXXX patterns that don't match, potentially causing incorrect issue counts.
Positive Aspects
TypeScript changes follow existing patterns with proper readonly modifiers and optional chaining
Critical-check prioritization uses stable sort to preserve input order within priority groups
Test coverage validates both positive and negative cases for the tags field
ADR-008 provides architectural context for the changes
The TypeScript implementation is solid. Address the shell script regex issue before merging.
You can now have the agent implement changes and create commits directly on your pull request's source branch. Simply comment with /q followed by your request in natural language to ask the agent to make changes.
The reason will be displayed to describe this comment to others. Learn more.
Pull Request Overview
The PR successfully implements the tags field and critical task prioritization, but several reliability issues in the new gc-watchdog skill and an algorithmic inefficiency in the runtime need to be addressed before merging.
While the project remains 'up to standards' according to Codacy, the gc-watchdog script is highly susceptible to crashes due to set -o pipefail interacting with grep assignments, and it lacks any automated test coverage. Additionally, the topological sort implementation has been changed from O(V+E) to O(V^2 log V). While acceptable for small plans, a more efficient two-queue approach should be used to maintain performance as the system scales.
About this PR
The gc-watchdog script relies heavily on regex/awk parsing of CLI tool outputs (gc status, bd list). This is inherently fragile as any change to the CLI output format will break the watchdog's health-check logic. Combined with the lack of automated tests for this script, this poses a high maintenance risk.
1 comment outside of the diffpackages/runtime/src/internal/topo.ts
line 16 ⚪ LOW RISK
Update the documentation at the top of topoSort to reflect that 'critical' tagged operations take priority over the default input order. The current comment still claims operations are emitted in input order, which is now only true within priority groups.
Test suggestions
Found: Validate that the IR PlanOperation accepts present, absent, and empty tags arrays.
Found: Verify that the SDK correctly copies tags from the specification to the generated plan.
Found: Verify that critical-tagged operations are sorted ahead of non-critical siblings when both are ready.
Found: Verify that topological sort preserves relative input order for operations within the same priority group.
Missing: Verify that the gc-watchdog script correctly identifies and filters 'wisp' and 'nudge' tasks from the open work count.
Missing: Verify that the gc-watchdog script exits with code 0 when the city is idle and healthy, but stays running on failures.
Prompt proposal for missing tests
Consider implementing these tests if applicable:
1. Missing: Verify that the gc-watchdog script correctly identifies and filters 'wisp' and 'nudge' tasks from the open work count.
2. Missing: Verify that the gc-watchdog script exits with code 0 when the city is idle and healthy, but stays running on failures.
• Adds optional tags?: readonly string[] to IR PlanOperation for prioritization, filtering, and
categorization.
• Passes spec.tags through from OperationSpec to PlanOperation in the SDK converter.
• Prioritizes "critical"-tagged operations among ready siblings in runtime topoSort via a new
stable sortReadyByPriority() helper.
• Adds ADR-008 documenting the decision, and a new project-local gc-watchdog skill (SKILL.md +
standalone script) for monitoring Gas City until idle.
• Adds TDD test coverage across IR validation, SDK conversion, and runtime topo sort.
Diagram
graph TD
A["OperationSpec.tags"] --> B["convertOperation()"] --> C["PlanOperation.tags"]
C --> D["topoSort()"]
D --> E["sortReadyByPriority()"]
E --> F["Ordered ready queue"]
G["gc-watchdog.sh"] --> H["gc status / bd list"]
Loading
High-Level Assessment
The chosen approach—an optional string-tags array with a simple stable 'critical first' sort in topoSort—is appropriate for the stated scope. It's additive (no schema bump), backward-compatible, and the ADR explicitly evaluates and rejects a boolean critical field and a full priority-queue scheduler as over-engineered for the current need. No better alternative stands out for this scale of requirement.
Files changed (9) +303 / -2
Enhancement (3) +38 / -2
plan.tsAdd optional tags field to PlanOperation+3/-0
Add optional tags field to PlanOperation
• Adds 'readonly tags?: readonly string[]' to the PlanOperation interface for prioritization, filtering, and categorization.
topo.tsPrioritize critical-tagged ops in topoSort ready queue+34/-2
Prioritize critical-tagged ops in topoSort ready queue
• Introduces sortReadyByPriority(), a stable sort placing ops tagged 'critical' first among ready siblings, applied both at initial ready-queue construction and whenever new operations become ready.
topo.test.tsAdd tests for critical-tag prioritization in topoSort+29/-0
Add tests for critical-tag prioritization in topoSort
• Adds tests confirming critical-tagged ops are scheduled first among independent ready ops and among siblings that become ready after a shared dependency completes.
ADR-008-tags-and-critical-prioritization.mdAdd ADR-008 for tags field and critical prioritization+60/-0
Add ADR-008 for tags field and critical prioritization
• Documents the decision to add tags to PlanOperation and prioritize critical-tagged ops in topoSort, including alternatives considered and consequences.
• New skill describing a periodic Gas City health monitor: check cycle, exit conditions, and reporting format for use by agents or as a standalone script.
• New bash script that polls gc status and bd list on an interval, reports mayor/session/controller health, and exits cleanly once no open or in-progress real work remains.
topoSort now calls op.tags?.includes("critical"), but validatePlan does not validate
operations[].tags at all. A malformed plan can pass IR validation and then throw at runtime when
scheduling/toposorting.
+ const isCritical = new Map<string, boolean>();+ for (const op of ops) {+ isCritical.set(op.id, op.tags?.includes("critical") ?? false);+ }
Evidence
The PR introduces runtime dependence on tags.includes(...) but IR validation still treats
operations as untyped records and enumerates per-op rules that do not include tags, allowing
invalid tags values to be considered valid plans and later crash runtime.
The issue below was found during a code review. Follow the provided context and guidance below and implement a solution
## Issue description
Runtime topoSort now evaluates `op.tags?.includes("critical")`. Optional chaining only protects `null|undefined`; if `tags` is present but not an array (e.g. `{}`), the method call can throw and crash scheduling.
## Issue Context
IR validation (`validatePlan`) is the boundary for untrusted plan JSON (via `deserializePlan`). It currently validates many per-operation fields but not `tags`.
## Fix Focus Areas
- packages/ir/src/validate.ts[170-220]
- packages/runtime/src/internal/topo.ts[37-41]
## Implementation notes
- In `validatePlan` per-operation loop, add a rule:
- If `op.tags` is defined, require `Array.isArray(op.tags)` and every element is a string.
- Add a `ValidationErrorDetail` with a new code (e.g. `INVALID_TAGS`) on failure.
- Optionally add a defensive check in `topoSort`:
- `const tags = op.tags; const critical = Array.isArray(tags) && tags.includes("critical");`
- Add/extend unit tests to cover invalid `tags` shapes (object/number elements) failing validation.
ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools
2. errexit breaks parsing 🐞 Bug☼ Reliability
Description
With set -euo pipefail, the unguarded grep|awk parsing pipelines for
MAYOR/SESSIONS/SUSPENDED/CONTROLLER can terminate the watchdog whenever a line is missing or format
changes. This contradicts the intended behavior of reporting issues and continuing to monitor on
failure.
+set -euo pipefail++INTERVAL="${1:-60}"+ITER=0++# Filter out ephemeral wisp/nudge beads from bd output.+# Real issues have IDs like sv-XXXX (4 chars after sv-).+# Wisp/nudge beads have IDs like sv-wisp-XXXX or sv-nudge-XXXX.+count_real_issues() {+ local status="$1"+ bd list --status="$status" 2>/dev/null \+ | grep -E '^\s*○ sv-[a-z0-9]{4} ' \+ | grep -v "wisp\|nudge" \+ | wc -l || true+}++while true; do+ ITER=$((ITER + 1))+ TS=$(date '+%H:%M:%S')++ # 1. gc status+ STATUS=$(timeout 15 gc status 2>/dev/null || true)++ if [ -z "$STATUS" ]; then+ echo "[$TS] #$ITER ⚠ gc status timeout/error"+ sleep "$INTERVAL"+ continue+ fi++ MAYOR=$(echo "$STATUS" | grep "harness.mayor" | awk '{print $2}')+ SESSIONS=$(echo "$STATUS" | grep "Sessions:" | head -1 | sed 's/^ *//')+ SUSPENDED=$(echo "$STATUS" | grep "Suspended:" | awk '{print $2}')+ CONTROLLER=$(echo "$STATUS" | grep "Controller:" | grep -o "supervisor-managed\|stopped\|error" | head -1)
Evidence
The script explicitly turns on errexit+pipefail and then performs several field extractions via
grep pipelines without || true. Since grep returns non-zero on no match, the watchdog can exit
immediately instead of emitting a warning and continuing.
The issue below was found during a code review. Follow the provided context and guidance below and implement a solution
## Issue description
The script enables `set -euo pipefail` but uses `grep` pipelines in command substitutions to extract fields. When `grep` finds no match it exits 1; with `pipefail` this fails the whole pipeline and with `-e` the script can exit immediately.
## Issue Context
This is especially likely during the abnormal situations the watchdog should keep running through (missing fields, partial output, format drift).
## Fix Focus Areas
- .agents/skills/gc-watchdog/watchdog.sh[11-55]
## Implementation notes
- Make each extraction non-fatal, e.g.:
- `MAYOR=$(echo "$STATUS" | awk '/harness\.mayor/{print $2; exit}' || true)`
- `SESSIONS=$(echo "$STATUS" | awk -F': ' '/Sessions:/{print $0; exit}' || true)`
- `SUSPENDED=$(echo "$STATUS" | awk '/Suspended:/{print $2; exit}' || true)`
- `CONTROLLER=$(echo "$STATUS" | awk '/Controller:/{print $2}' | grep -Eo '(supervisor-managed|stopped|error)' | head -1 || true)`
- Alternatively, temporarily disable errexit around parsing (`set +e` / `set -e`).
- Ensure downstream checks handle empty values (they already attempt to).
ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools
3. Watchdog miscounts work 🐞 Bug≡ Correctness
Description
watchdog.sh uses grep -E '^\s*…' to match leading whitespace, but \s is not POSIX ERE and is
not consistently supported across grep implementations. On affected systems this can undercount
open/in-progress work and incorrectly trigger the IDLE exit condition.
The script’s counting pipeline uses grep -E with \s, which is documented as not
meaningful/portable in POSIX grep regex modes; recommended portable form is [[:space:]]. This
directly affects the counts used in the exit condition.
The issue below was found during a code review. Follow the provided context and guidance below and implement a solution
## Issue description
The script relies on `grep -E '^\s*…'` to match leading whitespace. `\s` is not portable in POSIX ERE/BRE, so the match can fail depending on grep version/implementation, causing counts to be wrong (often 0) and the watchdog to exit early.
## Issue Context
This impacts the core exit condition logic: `OPEN_COUNT` / `INPROG_COUNT` can be computed incorrectly.
## Fix Focus Areas
- .agents/skills/gc-watchdog/watchdog.sh[19-25]
## Implementation notes
- Replace `^\s*` with `^[[:space:]]*` (or `^ *` if you only expect spaces).
- Consider switching the subsequent alternation uses to ERE too (e.g. `grep -E -v '(wisp|nudge)'`) for consistency.
- Keep the existing `|| true` guard in `count_real_issues`.
ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools
Review mode: ⚖️ Balanced: Behavioral changes span the IR/SDK contract, runtime scheduling, and an operational watchdog script, so a careful full review is warranted; the logic is not dense enough to justify redundant extended passes.
Tip of the day
💡 Did you know, you can group findings by type and pick your Finding display, from Minimal to Full
Backend-agnostic runtime: Executor interface (canExecute/execute), Scheduler
with topological sort, concurrent execution within maxConcurrent + optional
CPU/memory resource pool, dependent cancellation on fatal failure,
continueOnError, StateStore persistence for resume, CacheBackend reuse,
RetryPolicy, log/artifact collection.
<details>
- 64 vitest tests pass
- typecheck clean
- build green
- fixes sv-iqe: ResourcePool now created when either totalCpu or totalMemory
is set (independent optionals per spec), defaulting the unset dimension to
Infinity
</details>
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
…, version extraction
- Resolve relative --config/--baseline paths against --root (not process cwd)
via new internal/paths.ts resolveUnderRoot helper (baseline, execute, plan, validate)
- Serialize execute outcomes as plain object (Map→Object.fromEntries) for JSON output
- Forward --only-new to SDK in plan command (was silently dropped)
- Extract numeric version in doctor (regex x.y.z from "git version 2.43.0")
- Add 5s timeout to doctor runCheck (prevent hung shim blocking CLI)
- Wrap loadBaseline errors as CliError(SDK_ERROR) in baseline update/show
- Use process.exitCode instead of process.exit() in bin.ts (avoid truncating buffered I/O)
- Add tests: doctor version extraction, execute baseline path + outcomes, validate config path
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
… narrowing, function order
- init: use flag:"wx" (exclusive create) when !force to close TOCTOU race
between existsSync and writeFile; translate EEXIST to CONFIG_EXISTS
- bin.test.ts: fix it.skip() inside it() — tests silently passed when
dist/bin.mjs was missing; use it.skipIf(!binBuilt) at describe level
- doctor.test.ts: type-narrow JSON.parse result with `as` casts (no any)
- baseline.ts: reorder helper functions before dispatcher (readability)
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
bun.lock still recorded sverka bin as ./dist/index.mjs (the library
entry, no shebang/bootstrap) while package.json declares ./dist/bin.mjs.
Regenerated lockfile entry so lockfile-based installs link the correct
executable.
Addresses qodo-code-review and cubic-dev-ai review threads on bun.lock.
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Address Codacy finding: spawnSync without a timeout could block
indefinitely if a binary hangs. Added timeout: 5000 to match the
reliable behavior in doctor.ts.
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Pure function compileGithubWorkflow(plan, config?) that compiles a canonical
Sverka Plan to a GitHub Actions workflow YAML string. Thin wrapper per
ADR-004: single job runs `sverka execute .sverka/plan.json`. Maps plan
credentials to job-level env block with ${{ secrets.VAR }} references.
camelCase→kebab-case conversion for permissions and triggers. Deterministic
output. No custom errors (pure function on validated Plan). 13 tests pass.
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Add sitemap integration, canonical URLs, 404 page, robots.txt, and fix
all code examples to match the actual Sverka SDK API (defineWorkflow,
pipeline, task, run from @sverka/sdk). Remove references to non-existent
callback-based pipeline API and non-existent `sverka compile` CLI command.
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Address PR #17 review feedback (codeant-ai):
The quick start installed only the global CLI (@sverka/cli), but
sverka init generates a config that imports @sverka/sdk. Without
a project-local SDK install, plan/execute/validate fail with a
package resolution error. Added `bun add @sverka/sdk` to both
index.astro and getting-started.astro install instructions.
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Extract the four-role agent team (mayor/architect/builder/reviewer) and
workflow formulas (wave, address-review, bootstrap-sdd) from sverka-specific
root config into a reusable pack at pack/.
The pack is now imported under binding "harness" via [imports.harness] in
pack.toml. Project-specific context (project name, tech stack, wave plan) is
injected via append_fragments patches pointing to
template-fragments/project-context.md.
This separates three layers:
1. Harness (pack/): reusable roles + formulas + doc templates
2. Project docs (REVIEW.md, SECURITY.md, AGENTS.md): project-owned policy
3. Project content (specs/, engdocs/, packages/): project-owned code
To reuse in another project: import this pack via GitHub source, create a
project-context.md fragment, copy REVIEW.md/SECURITY.md/AGENTS.md templates.
Added:
- pack/ — sverka-gc-pack (agents, formulas, template-fragments, README)
- REVIEW.md — sverka review policy (two-axis, verification bar, commit hygiene)
- SECURITY.md — sverka security policy
- template-fragments/project-context.md — sverka context injected into agents
Changed:
- pack.toml — imports ./pack as harness, patches agents with project-context
- agents/{mayor,architect,builder,reviewer}/ — moved to pack/agents/
- formulas/sverka-{wave,bootstrap}.toml — moved to pack/formulas/ (renamed)
Verified: gc doctor clean, gc status shows harness.{mayor,architect,builder,
reviewer} agents loaded, formulas staged in .beads/formulas/.
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Add optional `tags: readonly string[]` to PlanOperation (IR) for
prioritization, filtering, and categorization. Pass tags through from
OperationSpec in convertToPlan (SDK). Prioritize critical-tagged ops
among ready siblings in topoSort (runtime).
Changes:
- ir/plan.ts: add `tags?: readonly string[]` to PlanOperation
- sdk/convert.ts: pass `spec.tags` to PlanOperation in convertOperation
- runtime/internal/topo.ts: sortReadyByPriority() — critical ops first
among ready siblings, then input order (stable)
Tests (TDD):
- ir: 3 tests — accepts tags, accepts no tags, accepts empty tags
- sdk: 2 tests — passes tags from spec, omits when absent
- runtime: 2 tests — critical first among independents, critical first
among siblings after dependency completes
Closes sv-bni (P3 backlog from Wave 3 / ADR-007).
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Records the architectural decision to add `tags` to PlanOperation and
prioritize critical-tagged ops in topoSort. Resolves the deferred
critical-check prioritization from ADR-007.
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Project-local skill at .agents/skills/gc-watchdog/ that monitors Gas City
health and exits when there is no more open or in-progress work.
- SKILL.md: documents check cycle, exit conditions, reporting format
- watchdog.sh: standalone script — checks gc status + bd list, filters
out ephemeral wisp/nudge beads, exits 0 when idle
Exit condition: no open real tasks AND no in-progress real tasks AND
mayor healthy. Does NOT exit on failure (only on success).
Usage: bash .agents/skills/gc-watchdog/watchdog.sh [interval_seconds]
Generated with [Devin](https://devin.ai)
Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Skipping CodeAnt AI review — this PR changes more than 100 files, which usually means a migration, codemod, or vendored drop. Line-level review on diffs this large produces duplicate findings on the same rewrite pattern and drowns out anything that actually matters.
If you still want a review, comment @codeant-ai : review. For better signal, consider splitting the PR into smaller chunks.
Closing this PR because it is now superseded by the main wave stack (#7–#25). The unique features in this branch can be reopened as focused PRs against the updated main once the main wave lands.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
codeant-ai
Bot
added
the
size:L
![amazon-q-developer[bot]](https://avatars.githubusercontent.com/in/1220912?s=60&v=4){kind=small}
baz-reviewer
Bot
added
baz: needs review
baz: pending
and removed
baz: needs review
labels
![codacy-production[bot]](https://avatars.githubusercontent.com/in/56611?s=60&v=4){kind=small}
![codeant-ai[bot]](https://avatars.githubusercontent.com/in/646884?s=60&v=4){kind=small}
![qodo-code-review[bot]](https://avatars.githubusercontent.com/in/484649?s=60&v=4){kind=small}
User description
Summary
Stacked on PR #19 (decouple Gas City harness). Three additions:
tagsfield on IRPlanOperation— optionalreadonly string[]for prioritization, filtering, categorizationtopoSort— critical-tagged ops scheduled first among ready siblings (stable sort, input order preserved within groups).agents/skills/gc-watchdog/with standalone script and exit conditionChanges
packages/ir/src/plan.tstags?: readonly string[]toPlanOperationpackages/sdk/src/convert.tsspec.tagsthrough inconvertOperationpackages/runtime/src/internal/topo.tssortReadyByPriority()— critical first among ready siblingsengdocs/adr/ADR-008-tags-and-critical-prioritization.md.agents/skills/gc-watchdog/SKILL.md.agents/skills/gc-watchdog/watchdog.shTests (TDD)
ir/validate.test.ts: 3 tests — accepts tags, accepts no tags, accepts empty tagssdk/convert.test.ts: 2 tests — passes tags from spec, omits when absentruntime/topo.test.ts: 2 tests — critical first among independents, critical first among siblings after dependencygc-watchdog exit condition
Exits when ALL true:
Does NOT exit on failure — only on clean idle.
Test plan
npx vitest run packages/ir/src/__tests__/validate.test.ts -t "tags"— 3 passnpx vitest run packages/sdk/src/__tests__/convert.test.ts -t "tags"— 2 passnpx vitest run packages/runtime/src/__tests__/topo.test.ts— 11 passnpx nx run ir, sdk, runtime: typecheck— cleannpx nx run ir, sdk, runtime: build— cleanbash .agents/skills/gc-watchdog/watchdog.sh 1— reports IDLE and exits 0Stacked on #19
Generated with Devin
Summary by cubic
Adds
tagsto IR operations, prioritizes"critical"checks among ready siblings, and adds agc-watchdogskill/script to exit cleanly when Gas City is idle. Also brings in CLI, checks, compilers, and docs; addresses Linear sv-bni by delivering the deferred critical-check prioritization from ADR-007.PlanOperation.tags?: readonly string[]; SDK propagatesspec.tags;topoSortschedules"critical"first (stable within groups)..agents/skills/gc-watchdog/withwatchdog.sh(filters wisp/nudge; exits only on clean idle with healthy mayor).@sverka/checks(builtin resolver + SARIF extraction),@sverka/cli(init/inspect/plan/execute/validate/baseline/doctor), compilers@sverka/compiler-githuband@sverka/compiler-gitlab.Written for commit 238bef0. Summary will update on new commits.
CodeAnt-AI Description
Add operation tags, prioritize critical work, and provide a Gas City watchdog
What Changed
criticalandsecurity, including empty tag lists.Impact
✅ Critical checks run sooner✅ Tags remain available for operation filtering and categorization✅ Clearer visibility into Gas City health and remaining work🔄 Retrigger CodeAnt AI Review
💡 Usage Guide
Checking Your Pull Request
Every time you make a pull request, our system automatically looks through it. We check for security issues, mistakes in how you're setting up your infrastructure, and common code problems. We do this to make sure your changes are solid and won't cause any trouble later.
Talking to CodeAnt AI
Got a question or need a hand with something in your pull request? You can easily get in touch with CodeAnt AI right here. Just type the following in a comment on your pull request, and replace "Your question here" with whatever you want to ask:
This lets you have a chat with CodeAnt AI about your pull request, making it easier to understand and improve your code.
Example
Preserve Org Learnings with CodeAnt
You can record team preferences so CodeAnt AI applies them in future reviews. Reply directly to the specific CodeAnt AI suggestion (in the same thread) and replace "Your feedback here" with your input:
This helps CodeAnt AI learn and adapt to your team's coding style and standards.
Example
Retrigger review
Ask CodeAnt AI to review the PR again, by typing:
Check Your Repository Health
To analyze the health of your code repository, visit our dashboard at https://app.codeant.ai. This tool helps you identify potential issues and areas for improvement in your codebase, ensuring your repository maintains high standards of code health.