Contexo is an open-source tool presented as “GitHub for AI agent knowledge, scoped per project,” using shared, git-backed knowledge so developers can switch among coding agents without repeatedly explaining project context. The supplied snippets establish that context continuity and usage budgets are real concerns across tools such as Claude Code and Cursor, but they provide no independent testing of Contexo’s transfer reliability, budget enforcement, or workflow friction; its maintainers are not identified in the supplied material.
Contexo repeats positions already established in Cognitive Git and Markdown OS—and already tracked through the Kungfu and Lethe portable-context cases—while adding Mole-like budget controls. With no independent reliability or friction evidence, it is another unvalidated implementation of known patterns rather than something that would yet change Scott’s architecture or claims.
ip:framework.cognitive-gitip:framework.markdown-osip:concept.autonomy-budgetdev:project.search-conversationsradar:kungfu-coding-agent-continuityradar:lethe-portable-ai-contextradar:mole-budgeted-research-agent
queries asked of Scott's wikis
- cross-agent context handoffs for coding agents
- git-backed agent memory and project knowledge
- portable memory across agent CLIs and IDEs
- inference budgets and hard cost controls for agents
- workflow friction in persistent coding-agent context
2026-08-23T12:33:50Z
No independent Contexo testing, adoption, or maintainer response emerged; the latest activity only repeats adjacent handoff risks and mitigations. As a low-relevance implementation of already-tracked patterns, this episode has faded and should be rediscovered only if direct usage evidence appears.
2026-08-21T12:29:19Z
A maintainer now claims an MCP-based Claude–Antigravity integration can be reliable, suggesting structured adapters and maintained instructions may mitigate generic handoff failures. This remains adjacent, self-reported evidence and does not test Contexo’s transfer fidelity, cost enforcement, or friction.
2026-08-21T01:24:44Z
The attached report supplies concrete failure evidence for an adjacent Claude-to-Antigravity delegation workflow, strengthening the general concern that cross-agent handoffs can lose constraints or introduce bad changes. It is not a Contexo test and provides no evidence on its budget controls, so the core hypothesis remains unvalidated.
2026-08-21T01:22:54Z
evidence attached: reddit.post.1vu1636 — The reported 6-of-9 delegation failures and ignored read-only instruction are concrete negative evidence about reliable cross-harness handoffs.
2026-08-20T08:35:11Z
Discussion now identifies two concrete handoff-quality requirements—detecting stale project state and preserving attempted-and-rejected approaches—in addition to reliable cost parsing. These sharpen the eventual test plan but remain hypothetical failure modes, with no independent use or implementation response to validate Contexo.
2026-08-18T08:27:17Z
A refreshed comment identifies a concrete weakness in cost-cap enforcement: inferring zero cost from CLI output could make a nominal hard cap silently advisory. This sharpens the validation target but is design criticism, not independent testing, so the case remains an unvalidated implementation of known patterns.
2026-08-18T06:46:26Z
The only change is negligible engagement growth, with no independent testing, adoption, or implementation evidence. Contexo remains an unvalidated example of already-known portable-context and budget-control patterns, so the case cools while awaiting real usage evidence.
2026-08-18T06:34:08Z
grounded: known/low — Contexo repeats positions already established in Cognitive Git and Markdown OS—and already tracked through the Kungfu and Lethe portable-context cases—while add
2026-08-18T06:31:17Z
origin walked (codex/luna, conf 0.94): anchor reddit.post.1vrgahm -> echo.github.bb0058b184 by sugihAF
2026-08-18T06:28:41Z
case created — The creator reports an Apache-licensed local CLI implementing cross-tool context handoffs and session cost governance.