Prompts are suggestions. Hooks are hard stops.
For months, I treated Claude Code like a junior developer who simply needed clear ground rules. My project instructions were explicit: never force-push, never delete branches, never run destructive commands. Most evenings, that worked. The agent wrote tests, refactored functions, and kept its hands off the git history. Then one rebase went sideways.
The context window filled with git error output. Conflict markers, detached HEAD messages, and branch divergence warnings stacked up, token by token. Buried under that noise was my polite instruction to avoid force-pushing. To the model, the most recent and salient text in the thread was the error stream. Statistical attention won over policy. The agent executed a command that wiped two hours of uncommitted local changes. It was not malicious; it was distracted. That distinction matters. An LLM does not break rules out of spite. It breaks them because a louder pattern in the context window temporarily overrides an earlier instruction.
That incident changed how I think about agent safety. A guardrail that works ninety-nine percent of the time is a liability. If the failure mode costs you time, money, or production data, you cannot leave it inside the prompt. You need enforcement outside the model’s reasoning loop.
Claude Code hooks solve exactly this. They are small scripts that intercept tool calls at three specific moments: before a tool executes (PreToolUse), after a tool finishes (PostToolUse), and when the agent decides it is done (Stop). Because they run as external code, they do not depend on the model’s memory, mood, or context pressure. The model can forget every instruction you ever gave it; the hook will still say no.
Here is the harness I built after that lost evening.
The Guard Hook: Intercept Before Damage
My PreToolUse hook inspects every Bash command before the shell touches it. I keep a tight denylist of destructive patterns. If the command string matches something dangerous, the hook aborts execution and returns an error directly back to the agent.
The patterns I block are simple and unambiguous:
git push --forceor any force-with-lease variant I do not trust yetgit reset --hardrm -rf
This is not sophisticated security research. It is a seatbelt. But the critical detail is what happens after the block.
I never return a blunt “Blocked.” A flat refusal confuses the agent and can trap it in a loop where it tries variations of the same destructive command. Instead, the error message includes an escape route. When the hook catches a hard reset, it tells the agent: “This command is blocked to protect uncommitted work. Commit a checkpoint first, then reassess.” That extra sentence changes the agent’s behavior completely. It pivots from attempting damage control to creating safety. The hook is not just a wall; it is traffic control.
I also chose a denylist over an allowlist for shell commands. At first, I considered allowing only an explicit set of safe git subcommands. That failed quickly. Agents are creatively literal. They run legitimate but unexpected commands like git stash push -m "wip" or git branch --show-current to check state. An allowlist breaks normal workflow the moment the model invents a valid but unlisted command. A short, curated denylist of genuinely destructive patterns gives the agent room to move while protecting the borders.
The Formatter Hook: Automate the Busywork
I used to waste prompt tokens telling the agent to “always run the formatter after editing a file.” It forgot half the time. The other half, it would pause and ask whether to format, burning a tool call on a decision that had only one right answer.
Now I handle that with a PostToolUse hook. After the agent edits a file, the hook checks the file extension. If it is Python, it runs Ruff. If it is JavaScript or TypeScript, it runs Prettier. If it is Go, it runs gofmt. The agent does not know the formatter exists. It does not need to.
Moving this out of the prompt had two effects. First, the code is consistently clean without adding cognitive load to the model. Second, my project instructions got shorter. Every “always” and “never” you remove from a prompt is a token the model can spend on actual problem-solving. The hook owns the invariant; the prompt owns the intent.
The Quality Gate: Redefining “Done”
The Stop hook runs when the agent decides it has finished the task and attempts to terminate the session. I do not let it. Instead, the hook runs the full test suite. If any test fails, the hook blocks the stop command and returns the failure output to the agent.
This changes the definition of completion. “Done” is no longer a feeling the model has. It is a measurable gate. The agent can only finish when the harness confirms the code works. In practice, this creates a tight feedback loop. The agent writes code, thinks it is finished, hits the stop button, and immediately sees a pytest traceback. It then self-corrects, fixes the import error or broken assertion, and tries to stop again. I have watched agents iterate three or four times inside this loop without human intervention. The harness enforces quality; the model supplies the patches.
What This Teaches About Agent Engineering
Building reliable autonomous systems requires a shift in mindset. You move from writing longer prompts to building tighter harnesses.
Use hooks for enforcement and prompts for policy. If a rule must hold one hundred percent of the time, it belongs in code, not in natural language. Prompts excel at ambiguity, taste, and architecture. They are terrible at invariants. If a mistake would cost you an afternoon of recovery time or, worse, production uptime, write a hook.
Shorter prompts produce better results. When you move mechanical rules into scripts, the model has less to remember and less to contradict. The agent’s context window is a scarce resource. Do not fill it with formatting reminders.
Finally, accept that your role is changing. As agents gain autonomy, the human’s job shifts from generating content to designing the guardrails. You are building the harness that decides what the model can touch, when it can finish, and how it must behave when things go wrong. That is engineering, not prompting.
The source that inspired this approach and additional implementation detail can be found here.
If you are building with AI agents and want to trade notes with other practitioners, you can find the GyaanSetu learning community here.