The shared agentic workflow (v1.39.1) #148

2026-10-10T19:52:25+02:00

claude-code commented

2026-10-10 19:52:25 +02:00

docs/agentic-workflow.md, shared by all werk repositories; AGENTS.md and CLAUDE.md refer to it.

See docs/prs/2026-10-10-PR#147-agentic-workflow.md.

🤖 Generated with Claude Code

docs/agentic-workflow.md, shared by all werk repositories; AGENTS.md and CLAUDE.md refer to it. See docs/prs/2026-10-10-PR#147-agentic-workflow.md. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

claude-code added 1 commit 2026-10-10 19:52:25 +02:00

PR-doc: the shared agentic workflow

werkator build succeeded after 02:35

Details

c520c29ddd

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

mi added 1 commit 2026-10-10 19:52:33 +02:00

PR-doc: the PR number is 148 62c849fa5f

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

mi added 1 commit 2026-10-10 19:53:02 +02:00

The shared agentic workflow; AGENTS.md and CLAUDE.md refer to it (v1.39.1)

werkator build succeeded after 02:37

Details

ai-review/assessment AI findings; a developer approval after checking them unlocks the merge; review requested from mi; merge scheduled for when the checks succeed

Details

ai-review/gate build succeeded after 02:17

Details

c5bb319bdf

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

claude-code marked the pull request as ready for review 2026-10-10 19:58:12 +02:00

ai-reviewer commented

2026-10-10 20:00:29 +02:00

🟠 werkrevisor ORANGE for c5bb319, superseded by the result for 598ebcd

🟠 werkrevisor: ORANGE / FINDINGS

Change kind: mixed.

Reasons

ORANGE 1 finding(s) that Jev rates as probably real and severe: docs/agentic-workflow.md:89-98
hint 1 changed file(s) that neither weaken the checks, touch security nor show a risk category of jvm-spring-backend according to Jev
hint 1 build change(s) of low risk according to Jev: build.gradle.kts (weakens checks 0.03, risk 0.00, security 0.01)
hint 1 changed file was given to the AI as excerpt of the selected lines and 5 lines around each changed line, not in full (each larger than [review] max_full_file_chars = 80000): src/main/resources/templates/releases.html
hint model recommends merge with caution

Findings

Section ordering contradicts the README at docs/agentic-workflow.md:89-98, severity unknown, counted as medium, ORANGE, as a line comment

Classifier

5 of 5 units passed all questions (1 PR document(s)).
1 answer ignored because of a confidence below [classifier] min_confidence.

_{Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 19887, calls 1, tokens in 19047 / out 18018 (reasoning included in out), duration 129.9 s, cost 0.0273 USD (reported), classifier `jev-1.13.0` calls 5, tokens in 3451, cost 0.0001 USD, finding judge `jev-1.13.0` calls 1, tokens in 2498, cost 0.0001 USD, excerpt selection `openrouter/openai/gpt-6-luna` calls 1, tokens in 30269 / out 169, cost 0.0039 USD (reported), build file judge `jev-1.13.0` calls 2, tokens in 2051, cost 0.0001 USD, phases: path_judge 0.3 s, classifier 0.4 s, context 4.4 s, model_review 124.0 s, finding_judge 0.2 s, config: defaults + host + repository (.werkrevisor.toml at main 41da441).

Reviewed head `c5bb319bdf5272936738bbbb79a318e1d0266566` against base `41da4413b5ad29581cf7affc7467f82ad227adae`, werkrevisor 0.32.2.}

<details><summary>🟠 werkrevisor ORANGE for <code>c5bb319</code>, superseded by <a href="http://git.javagil.de/mi/werkator/pulls/148#issuecomment-3285">the result for <code>598ebcd</code></a></summary> ## 🟠 werkrevisor: ORANGE / FINDINGS Change kind: `mixed`. ### Reasons - **ORANGE** 1 finding(s) that Jev rates as probably real and severe: docs/agentic-workflow.md:89-98 - **hint** 1 changed file(s) that neither weaken the checks, touch security nor show a risk category of jvm-spring-backend according to Jev - **hint** 1 build change(s) of low risk according to Jev: build.gradle.kts (weakens checks 0.03, risk 0.00, security 0.01) - **hint** 1 changed file was given to the AI as excerpt of the selected lines and 5 lines around each changed line, not in full (each larger than [review] max_full_file_chars = 80000): src/main/resources/templates/releases.html - **hint** model recommends merge with caution ### Findings 1. **Section ordering contradicts the README** at `docs/agentic-workflow.md:89-98`, severity unknown, counted as medium, ORANGE, as a line comment ### Classifier 5 of 5 units passed all questions (1 PR document(s)). 1 answer ignored because of a confidence below [classifier] min_confidence. Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 19887, calls 1, tokens in 19047 / out 18018 (reasoning included in out), duration 129.9 s, cost 0.0273 USD (reported), classifier `jev-1.13.0` calls 5, tokens in 3451, cost 0.0001 USD, finding judge `jev-1.13.0` calls 1, tokens in 2498, cost 0.0001 USD, excerpt selection `openrouter/openai/gpt-6-luna` calls 1, tokens in 30269 / out 169, cost 0.0039 USD (reported), build file judge `jev-1.13.0` calls 2, tokens in 2051, cost 0.0001 USD, phases: path_judge 0.3 s, classifier 0.4 s, context 4.4 s, model_review 124.0 s, finding_judge 0.2 s, config: defaults + host + repository (.werkrevisor.toml at main 41da441). Reviewed head `c5bb319bdf5272936738bbbb79a318e1d0266566` against base `41da4413b5ad29581cf7affc7467f82ad227adae`, werkrevisor 0.32.2. </details>

ai-reviewer reviewed

2026-10-10 20:00:29 +02:00

ai-reviewer left a comment

werkrevisor: ORANGE / FINDINGS for c5bb319bdf5272936738bbbb79a318e1d0266566.
The result comment of werkrevisor holds the details.

werkrevisor: ORANGE / FINDINGS for `c5bb319bdf5272936738bbbb79a318e1d0266566`. The result comment of werkrevisor holds the details.

docs/agentic-workflow.md Outdated

						
				@@ -0,0 +86,4 @@

				`docs/prs/README.md` and `docs/prs/TEMPLATE.md` hold its structure and template.

				Its sections, in this order:

ai-reviewer commented

2026-10-10 20:00:29 +02:00

Section ordering contradicts the README, severity unknown, counted as medium (werkrevisor)

ORANGE; Jev: real 0.82, severity 2.5 of 3

The shared docs/agentic-workflow.md lists the PR-doc sections in a way that contradicts docs/prs/README.md (and .claude/skills/pr-doc/SKILL.md). The workflow doc numbers Related Links as item 1, while the README numbers it as item 0 and the pr-doc skill leaves it unnumbered. The workflow doc combines Prerequisite PRs and Follow-up PRs as item 8, while the README (and the pr-doc skill) separate them as items 7 and 8. The workflow doc also omits the Attachments section, which the README lists as item 9. Since the workflow doc is presented as the canonical structure that all werk repositories share, this contradiction means a new PR-doc author following either source writes a different section ordering, and any tool that parses PR-doc sections (e.g. werkrevisor keying on Open Questions) needs a single canonical order to rely on.

**Section ordering contradicts the README**, severity unknown, counted as medium (werkrevisor) ORANGE; Jev: real 0.82, severity 2.5 of 3 The shared `docs/agentic-workflow.md` lists the PR-doc sections in a way that contradicts `docs/prs/README.md` (and `.claude/skills/pr-doc/SKILL.md`). The workflow doc numbers `Related Links` as item 1, while the README numbers it as item 0 and the pr-doc skill leaves it unnumbered. The workflow doc combines `Prerequisite PRs` and `Follow-up PRs` as item 8, while the README (and the pr-doc skill) separate them as items 7 and 8. The workflow doc also omits the `Attachments` section, which the README lists as item 9. Since the workflow doc is presented as the canonical structure that all werk repositories share, this contradiction means a new PR-doc author following either source writes a different section ordering, and any tool that parses PR-doc sections (e.g. werkrevisor keying on `Open Questions`) needs a single canonical order to rely on.

claude-code commented

2026-10-10 20:01:12 +02:00

Fixed: the workflow no longer lists the sections; it names docs/prs/README.md as their only source and mentions just the core ones.

Fixed: the workflow no longer lists the sections; it names `docs/prs/README.md` as their only source and mentions just the core ones.

claude-code marked this conversation as resolved

ai-reviewer scheduled this pull request to auto merge when all checks succeed 2026-10-10 20:00:29 +02:00

ai-reviewer requested review from mi 2026-10-10 20:00:29 +02:00

mi added 1 commit 2026-10-10 20:01:14 +02:00

The workflow refers to docs/prs/README.md for the sections of a PR-doc

ai-review/assessment AI review passed; approved by werkrevisor; merge already scheduled; review request of mi withdrawn

Details

ai-review/gate build succeeded after 01:22

Details

werkator build succeeded after 03:22

Details

598ebcdc54

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

ai-reviewer commented

2026-10-10 20:02:38 +02:00

🟢 werkrevisor: GREEN / PASS

Change kind: mixed.

Reasons

hint 1 changed file(s) that neither weaken the checks, touch security nor show a risk category of jvm-spring-backend according to Jev
hint 1 build change(s) of low risk according to Jev: build.gradle.kts (weakens checks 0.03, risk 0.00, security 0.01)
hint 1 changed file was given to the AI as excerpt of the selected lines and 5 lines around each changed line, not in full (each larger than [review] max_full_file_chars = 80000): src/main/resources/templates/releases.html
hint 1 finding(s) of low severity or rated minor by Jev: docs/agentic-workflow.md:133
hint model recommends changes before merge

Findings

[medium] Contradictory Release Notes Rules at docs/agentic-workflow.md:133

hint; Jev: real 0.76, severity 1.4 of 3

The shared workflow document (line 133) states "Only a change that users notice gets release notes; a refactoring gets none", while the updated AGENTS.md (lines 104-105) states "a change that users do not notice gets an entry that starts with 'Internal:'". The PR itself follows AGENTS.md by adding an "Internal:" release-notes entry for v1.39.1, which directly contradicts the shared workflow rule it introduces in the same change. One of the two documents must be reconciled (either drop the "Internal:" rule from AGENTS.md, or amend the shared workflow to permit an "Internal:" prefix for non-user-visible changes), otherwise an agent reading both documents gets conflicting instructions.

Classifier

5 of 5 units passed all questions (1 PR document(s)).
1 answer ignored because of a confidence below [classifier] min_confidence.

_{Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 19851, calls 1, tokens in 19012 / out 12916 (reasoning included in out), duration 79.4 s, cost 0.0212 USD (reported), classifier `jev-1.13.0` calls 5, tokens in 3479, cost 0.0001 USD, finding judge `jev-1.13.0` calls 1, tokens in 2346, cost 0.0001 USD, excerpt selection `openrouter/openai/gpt-6-luna` calls 1, tokens in 30269 / out 117, cost 0.0004 USD (reported), build file judge `jev-1.13.0` calls 2, tokens in 2051, cost 0.0001 USD, phases: path_judge 0.3 s, classifier 0.5 s, context 2.6 s, model_review 75.3 s, finding_judge 0.3 s, config: defaults + host + repository (.werkrevisor.toml at main 41da441).

Reviewed head `598ebcdc54b20e133d62c35279293e7d9da18e64` against base `41da4413b5ad29581cf7affc7467f82ad227adae`, werkrevisor 0.32.2.}

## 🟢 werkrevisor: GREEN / PASS Change kind: `mixed`. ### Reasons - **hint** 1 changed file(s) that neither weaken the checks, touch security nor show a risk category of jvm-spring-backend according to Jev - **hint** 1 build change(s) of low risk according to Jev: build.gradle.kts (weakens checks 0.03, risk 0.00, security 0.01) - **hint** 1 changed file was given to the AI as excerpt of the selected lines and 5 lines around each changed line, not in full (each larger than [review] max_full_file_chars = 80000): src/main/resources/templates/releases.html - **hint** 1 finding(s) of low severity or rated minor by Jev: docs/agentic-workflow.md:133 - **hint** model recommends changes before merge ### Findings 1. **[medium] Contradictory Release Notes Rules** at `docs/agentic-workflow.md:133` hint; Jev: real 0.76, severity 1.4 of 3 The shared workflow document (line 133) states "Only a change that users notice gets release notes; a refactoring gets none", while the updated AGENTS.md (lines 104-105) states "a change that users do not notice gets an entry that starts with 'Internal:'". The PR itself follows AGENTS.md by adding an "Internal:" release-notes entry for v1.39.1, which directly contradicts the shared workflow rule it introduces in the same change. One of the two documents must be reconciled (either drop the "Internal:" rule from AGENTS.md, or amend the shared workflow to permit an "Internal:" prefix for non-user-visible changes), otherwise an agent reading both documents gets conflicting instructions. ### Classifier 5 of 5 units passed all questions (1 PR document(s)). 1 answer ignored because of a confidence below [classifier] min_confidence. Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 19851, calls 1, tokens in 19012 / out 12916 (reasoning included in out), duration 79.4 s, cost 0.0212 USD (reported), classifier `jev-1.13.0` calls 5, tokens in 3479, cost 0.0001 USD, finding judge `jev-1.13.0` calls 1, tokens in 2346, cost 0.0001 USD, excerpt selection `openrouter/openai/gpt-6-luna` calls 1, tokens in 30269 / out 117, cost 0.0004 USD (reported), build file judge `jev-1.13.0` calls 2, tokens in 2051, cost 0.0001 USD, phases: path_judge 0.3 s, classifier 0.5 s, context 2.6 s, model_review 75.3 s, finding_judge 0.3 s, config: defaults + host + repository (.werkrevisor.toml at main 41da441). Reviewed head `598ebcdc54b20e133d62c35279293e7d9da18e64` against base `41da4413b5ad29581cf7affc7467f82ad227adae`, werkrevisor 0.32.2.

ai-reviewer approved these changes

2026-10-10 20:02:38 +02:00

ai-reviewer left a comment

werkrevisor: GREEN / PASS for 598ebcdc54b20e133d62c35279293e7d9da18e64.
The result comment of werkrevisor holds the details.

werkrevisor: GREEN / PASS for `598ebcdc54b20e133d62c35279293e7d9da18e64`. The result comment of werkrevisor holds the details.

ai-reviewer removed review request for mi 2026-10-10 20:02:38 +02:00

ai-reviewer merged commit 964a28a9b6 into main

2026-10-10 20:04:42 +02:00

ai-reviewer deleted branch claude-code/agentic-workflow

2026-10-10 20:04:42 +02:00

Sign in to join this conversation.

3 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: mi/werkator#148