Agentic Context Engine is an open-source engine that gives AI agents a persistent learning loop, helping them remember successful strategies and learn from failures across sessions. It is used to improve production agents, and also powers Kayba’s hosted service. Catalogue add-ons support workflows for operating and configuring the engine.
Borrowing it
Nothing to install: this file belongs to kayba-ai/agentic-context-engine. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/kayba-ai/agentic-context-engine/main/.claude/commands/finalize.mdgit clone --depth 1 https://github.com/kayba-ai/agentic-context-engineWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/kayba-ai/agentic-context-engine/finalize)<a href="https://agentmods.dev/commands/kayba-ai/agentic-context-engine/finalize"><img src="https://agentmods.dev/badge/commands/kayba-ai/agentic-context-engine/finalize/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/kayba-ai/agentic-context-engine/finalize"><img src="https://agentmods.dev/badge/commands/kayba-ai/agentic-context-engine/finalize.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.00902 |
| Opus 5.5 | $0.00000 | $0.00361 |
| Sonnet 5.5 | $0.00000 | $0.00180 |
| Haiku 4.5 | $0.00000 | $0.00090 |
Grade A, and why
finalize scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Finalize the current work: format, test, fix, commit, and push.
Run this command after making code changes to complete the development cycle. It handles formatting, testing (with auto-fix retries), committing, and pushing.
Arguments: $ARGUMENTS is an optional commit message override. If not provided, compose one automatically from the diff.
Workflow:
-
Format code
- Run
uv run black ace/ tests/ examples/
- Run
-
Update CLAUDE.md
- Run
/initto update the project's CLAUDE.md file with current codebase context
- Run
-
Review test coverage
- Check if the changes have adequate test coverage
- Look at which lines/branches are untested for the modified files
- If coverage gaps exist for the changed code, add targeted tests before proceeding
- Focus on: new functions, error paths, edge cases, and branches introduced by this changeset
-
Run tests
- Run
uv run pytest
- Run
-
Test-fix loop (max 3 retries)
- If tests pass, continue to step 6
- If tests fail:
- Analyze the failure output
- Fix the failing code or tests
- Add missing tests if the failures reveal gaps
- Re-run formatter:
uv run black ace/ tests/ examples/ - Re-run tests:
uv run pytest - If tests still fail after 3 total attempts, stop entirely and report the failures. Never commit broken code.
-
Security review
- Run
/security-reviewto scan changed code for vulnerabilities - If Critical or High severity issues are found, fix them before proceeding
- After fixing, re-run formatter (
uv run black ace/ tests/ examples/) and tests (uv run pytest) - If issues can't be auto-fixed, report them and stop
- Run
-
Review changes
- Run
git diffto review all changes - Run
git statusto see untracked/modified files - Determine which files to stage
- Run
-
Stage files selectively
- Use explicit
git add <file>for each file. Never usegit add -Aorgit add . - Never stage:
.env, credentials, secrets,__pycache__/,*.pyc, large binaries,.DS_Store - Only stage
uv.lockif dependency changes inpyproject.tomlwere intentional - If unsure about a file, ask the user
- Use explicit
-
Compose commit message
- If $ARGUMENTS was provided, use it as the commit message
- Otherwise, compose a Conventional Commit message:
<type>(<scope>): <short description> - Types:
feat,fix,refactor,test,docs,chore,perf,style - Derive scope from the primary changed file path (e.g.,
ace/skillbook.py->skillbook,ace/integrations/litellm.py->integrations,tests/test_foo.py->tests) - Keep messages short and imperative
-
Check branch safety
- Run
git branch --show-currentto get the current branch - If on
mainormaster, warn the user and ask for explicit confirmation before committing - If denied, stop without committing
- Run
-
Commit and push
- Commit with the composed message
- Push to remote:
git push(orgit push -u origin <branch>if no upstream is set)
-
Report summary
- Commit hash (short)
- Branch name
- Files changed count
- Test results (pass count)
- Any warnings encountered
Error handling:
- If no changes exist (clean working tree), report "Nothing to finalize" and stop
- If tests fail after 3 retries, report failures and stop without committing
- If push fails, report the error but keep the local commit
- If formatter fails, report the error and stop
- If security review finds unfixable Critical/High issues, report and stop without committing
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 80 lines · 0 tokens per session scan A 10b4ca0fefdc
finalize is a command published in the GitHub repository kayba-ai/agentic-context-engine (2,608 stars, last pushed 1mo ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 902 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
finalize compared with commands from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| finalizekayba-ai/agentic-context-engine | Command | 2.6k | 1mo ago | Apache-2.0 | A | 0 | Claude Code |
| dream-applyconorbronsdon/agent-memory-kit | Command | 2 | 1mo ago | MIT | A | 20 | Claude Code |
| endconorbronsdon/agent-memory-kit | Command | 2 | 1mo ago | MIT | A | 14 | Claude Code |
| dreamconorbronsdon/agent-memory-kit | Command | 2 | 1mo ago | MIT | A | 0 | Claude Code |
| startconorbronsdon/agent-memory-kit | Command | 2 | 1mo ago | MIT | A | 16 | Claude Code |
| updateconorbronsdon/agent-memory-kit | Command | 2 | 1mo ago | MIT | A | 12 | Claude Code |
| gsd:undomrboups/xbrain | Command | 2 | 1mo ago | MIT | A | 23 | Claude Code |