[ 
https://issues.apache.org/jira/browse/CAMEL-26960?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18144746#comment-18144746
 ] 

Claus Ibsen commented on CAMEL-26960:
-------------------------------------

Step 1 done: commit 3d34c81fcef9 on the local branch CAMEL-26960-answer-marks, 
not pushed yet. Claus chose marks, no repair turn.

*What is checked.* The code blocks of an answer, in the AI panel and in camel 
ask:
* YAML: unknown keys and options, from the YAML DSL schema. A fragment is 
wrapped in what it belongs to: a {{from:}} without its route goes in a route 
file, a list of steps and a single EIP go in the steps of a route, and 
{{redeliveryPolicy:}}, {{resilience4jConfiguration:}} and {{onFallback:}} go 
under their EIP.
* Endpoint uris: options the component does not have.
* {{camel.*}} keys of properties lines, in a block of any language (the model 
writes them in yaml blocks too).
* A value of the wrong type, only when the validator has a hint for it, such as 
{{onException}} written as a list with the exception class as a key.

*What is not marked.* A required option left out, a placeholder, a bare "string 
found, object expected", the options of an EIP shown without the EIP 
({{slidingWindowSize: 10}}), and keys of other runtimes. Code in an answer is 
partial on purpose, and a wrong mark costs trust.

*Where the marks show.*
* The AI panel: a ✗ line under the block, with the line in the block and the 
hint of the validator. Copy still copies only the code.
* camel ask: a {{Check:}} line on stderr, so stdout is still only the answer.
* The model is told of the marks, and of a simple expression it left invalid, 
with the next question. The answer in the history stays as the model wrote it.

*Measured on the benchmark answers of Nov 13* (q1 + q1b, 318 answers, 305 code 
blocks): 53 marks in 27 answers, and each one is a real mistake (read one by 
one). The whole run took 0.75 s, validator start included.

*Two validator bugs, found by that run and fixed in the same commit:*
* A commented-out option under {{parameters:}} was reported as the unknown 
option '# idempotentRepository'. This also affects camel_validate_source.
* A bean with its {{name:}} and an unknown key ({{parameters:}}) got the hint 
"the name is a property, not the key". It now gets "goes under properties: or 
constructors:".

*Left for the next steps:*
* Option maps without their key ({{slidingWindowSize: 10}}, 6 blocks), which 
needs the definition index of step 2.
* Unknown {{camel.*}} groups ({{camel.tracing.enabled}}, {{camel.profile}}). 
The catalog does not know them, but it does not know 
{{camel.servlet.mapping.context-path}} of Spring Boot either, so they need care.
* Parse errors: their line and column numbers point into the wrapped block.
* ACP agents, and the markdown mode of camel_validate_source.

_Claude Code on behalf of Claus Ibsen (davsclaus)_


> camel-jbang - Mark the mistakes in the code of AI answers, such as unknown 
> options
> ----------------------------------------------------------------------------------
>
>                 Key: CAMEL-26960
>                 URL: https://issues.apache.org/jira/browse/CAMEL-26960
>             Project: Camel
>          Issue Type: Improvement
>          Components: camel-jbang
>            Reporter: Claus Ibsen
>            Assignee: Claus Ibsen
>            Priority: Major
>             Fix For: 4.24.0
>
>
> The AI clients of the Camel CLI show code in their answers: YAML routes, 
> application.properties lines, endpoint URIs. Nothing checks that code. A file 
> the model writes goes through the validator, but code in an answer does not, 
> and a beginner copies it.
> *Data* from a benchmark with a local model (Nov 13-14 2026, qwen3.6:35b-mlx, 
> 5 example apps, 53 beginner questions in 3 rounds, so 159 answers per run):
> * Without an instruction: 6 unknown options in 4 answers.
> * With an instruction in the first message to look options up and to check 
> every snippet with camel_validate_source: 32 unknown options in 14 answers. 
> The model wrote 221 code blocks (84 without the instruction) and checked none 
> of them. Its 3 calls of camel_validate_source checked the route file, not a 
> snippet.
> Examples from the answers:
> * {{idempotentConsumer}} with {{messageIdCache}} and {{cacheName}} (the 
> option is {{idempotentRepository}})
> * {{redeliveryPolicy}} with {{multiplier}} (the option is 
> {{backOffMultiplier}})
> * {{camel.metrics-enabled}} (the key is {{camel.metrics.enabled}}), 
> {{camel.tracing}}, {{camel.health.check-routes-on-startup}}
> * {{log:ERROR?level=ERROR&message=...}} (the log endpoint has no message 
> option)
> * {{onException}} with the exception class as a YAML key
> * {{forceOpen}}, {{failureThreshold}} and {{openDuration}} on a circuit 
> breaker
> *Proposal: mark the mistakes, do not change the answer.*
> * When an answer is complete, take its fenced code blocks (YAML, XML, Java, 
> properties) and check them with the same checks as camel_validate_source with 
> content.
> * Under each block with a problem, show one line per problem with the nearest 
> valid names. For example: "multiplier is not an option of redeliveryPolicy; 
> did you mean backOffMultiplier?"
> * Add the findings to the conversation, so the next answer does not repeat 
> the mistake.
> * No extra model call. Always on.
> *Where*
> * The check itself goes in camel-jbang-core, next to camel_validate_source, 
> so all clients share it.
> * camel ask: print the findings below the answer.
> * The AI panel: mark the blocks in the answer, for the built-in model and for 
> ACP agents (their answers come in the same way).
> * External MCP clients never show Camel their answer. They can get a tool 
> that checks the code blocks of a text. This can be a follow-up.
> *Fragments*
> Most code in an answer is a fragment, not a whole route. {{camel validate 
> yaml}} finds {{messageIdCache}} in a whole route. But it refuses a bare 
> {{redeliveryPolicy:}} block for its shape ("object found, array expected") 
> before it checks the options. The check needs a fragment mode: check the 
> block against the schema of its EIP, or wrap a list of steps in a route.
> *How to measure*
> The benchmark counts unknown options in answers. With this in place, no 
> unknown option should be shown without a mark (32 in the run with the 
> instruction).
> *Later, not in this issue:* a repair turn, where the model gets the findings 
> and corrects its answer.



--
This message was sent by Atlassian Jira
(v10.3.26#10030026)

Reply via email to