- I've seen Claude ignore important parts of skills/agent files multiple times. I was running a clean up SKILL.md on a hundred markdown files, manually in small groups of 5, and about half the time it listened and ran the skill as written. The other half it would start trying to understand the codebase looking for markdown stuff for 2min, for no good reason, before reverting back to what the skill said.
LLMs are far from consistent.
- For those unaware, Claude Code comes with a built in /insights command...
- > 26% of sessions are abandoned, most within the first 60 seconds
Starting new sessions frequently and using separate new sessions for small tasks is a good practice.
Keeping context clean and focused is a highly effective way to keep the agent on task. Having an up to date AGENTS.md should allow for new sessions to get into simple tasks quickly so you can use single-purpose sessions for small tasks without carrying the baggage of a long past context into them.
by monsterxx03
0 subcomment
- I built something in a similar space: Linko (https://github.com/monsterxx03/linko), a transparent MITM proxy with a webui that lets you see what's actually being sent between Claude Code and LLM APIs in
real time.
It's been really helpful for me to debug my own sessions and understand what the model is seeing (system prompts, tool definitions, tracing tool calls etc.).
- The "no meaningful benchmark for good agentic session performance" point resonates. Success varies so much by task type that a single metric is almost meaningless. A 60-second documentation lookup and a 30-minute refactoring session could both be successes.
Curious what shape the benchmark takes. Are you thinking per-task-type baselines, or something more like an aggregate efficiency score?
by steve_adams_86
0 subcomment
- Does this comply with Anthropic's terms? I've been developing small apps here and there on top of Claude Code and each time I find I'm too uncomfortable with their terms to bother distributing it.
by swaminarayan
2 subcomments
- 26% of AI coding sessions are abandoned within 60 seconds .Is this a prompt problem, a tooling problem, or a limitation of current models?
- I have seen numbers claiming tools are only called 59% of the time.
Saw another comment on a different platform where someone floated the idea of dynamically injecting context with hooks in the workflow to make things more deterministic.
- I see a lot of people with concerns about privacy and security. Not shown in the post, but the github shows how to self host. No need to use 3rd party, you can just have your own too
- is there a reason, other than general faith in humanity, to assume those '1573 sessions' are real?
I do not see any link or source for the data. I assume it is to remain closed, if it exists.
by marconardus
1 subcomments
- It might be worthwhile to include some of an example run in your readme.
I scrolled through and didn’t see enough to justify installing and running a thing
- Reminds me https://www.agentsview.io/.
by KaiserPister
1 subcomments
- This is awesome! I’m working on the Open Prompt Initiative as a way for open source to share prompting knowledge.
- Why does it need login and cloud upload? A local cli tool analyzing logs should be sufficient.
- So what conclusions have you drawn or could a person reasonably draw with this data?
by ekropotin
1 subcomments
- > That's it. Your Claude Code sessions will now be uploaded automatically.
No, thanks
by smallerfish
1 subcomments
- > content, the content or transcript of the agent session
Does this include the files being worked on by the agent in the session, or just the chat transcript?
- I 100% agree that we need tools to understand and audit these workflows for opportunities. Nice work.
TBH, I am very hesitant to upload my CC logs to a third-party service.
by anthonySs
1 subcomments
- is this observability for your claude code calls or specifically for high level insights like skill usage?
would love to know your actual day to day use case for what you built
by mentalgear
1 subcomments
- How diverse is your dataset?
- Why is the comment calling out the biggest issue with this so heavily downvoted? Privacy is a massive concern with this.
- Does it work for Codex?
- One potential reason for sessions being abandoned within 60 seconds in my experience is realizing you forgot to set something in the environment: github token missing, tool set for the language not on the path, etc. Claude doesn't provide elegant ways to fix those things in-session so I'll just exit, fix up and start Claude again. It does have the option to continue a previous session but there's typically no point in these "oops I forgot that" cases.
by cluckindan
2 subcomments
- Nice. Now, to vibe myself a locally hosted alternative.
by sriramgonella
1 subcomments
- [flagged]
- [flagged]
by socialinteldev
2 subcomments
- [flagged]
- [dead]
by longtermemory
1 subcomments
- From session analysis, it would be interesting to understand how crucial the documentation, the level of detail in CLAUDE.md, is.
It seems to me that sometimes documentation (that's too long and often out of date) contributes to greater entropy rather than greater efficiency of the model and agent.
It seems to me that sometimes it's better and more effective to remove, clean up, and simplify (both from CLAUDE.md and the code) rather than having everything documented in detail.
Therefore, from session analysis, it would be interesting to identify the relationship between documentation in CLAUDE.md and model efficiency. How often does the developer reject the LLM output in relation to the level of detail in CLAUDE.md?
by aplomb1026
0 subcomment
- [flagged]
- [dead]
by Sebastian_Dev
0 subcomment
- [dead]
by huflungdung
0 subcomment
- [dead]
- [flagged]
by multidude
2 subcomments
- [flagged]
by mihir_kanzariya
2 subcomments
- [flagged]
- [flagged]
by ozgurozkan
2 subcomments
- [flagged]
- This is so sad that on top of black box LLMs we also build all these tools that are pretty much black box as well.
It became very hard to understand what exactly is sent to LLM as input/context and how exactly is the output processed.
Hacker News