git:20260822.948d3ff to git:20260822.ac50607

---

name: delete-guard

description: >-

Recoverable deletion discipline for AI agents. Use whenever you need to

delete, clean, or discard files/directories or run destructive git

operations (clean, reset --hard, restore, force push) inside a workspace.

Guarantees quarantine-with-manifest instead of irreversible loss, and

explains what to do when the guard blocks you.

---

# delete-guard

You operate inside a workspace where irreversible destruction is not a

default capability. This skill defines the deletion discipline the guard

enforces and how to work with it productively. The four pillars:

- **Scope** - you act inside the workspace; the workspace root, `.git`, and

everything outside are never yours to delete.

- **Recoverability** - deletions are compensated (relocated to `.agent-trash/`

with a manifest, or snapshotted as a git stash) before they take effect.

- **Authorization** - a human decides permission boundaries. You can be

*downgraded* to RESTRICTED mode; you can never promote yourself back.

- **Auditability** - every verdict, relocation, and restore is recorded in

JSONL. Nothing is invisible.

## The one rule

**Prefer `safe_delete` over `rm`. It is not slower for you - it is the

supported path.**

```bash

# delete files/dirs/globs - they are quarantined, not destroyed:

python3 <repo>/skills/delete-guard/scripts/safe_delete.py src/old_module.py

python3 <repo>/skills/delete-guard/scripts/safe_delete.py 'build/**/*.tmp' --reason "stale build output"

# undo a transaction (list first if unsure):

python3 <repo>/skills/delete-guard/scripts/restore.py list

python3 <repo>/skills/delete-guard/scripts/restore.py <txid>

# current guard state:

python3 <repo>/skills/delete-guard/scripts/status.py

```

Globs passed to `safe_delete` are expanded explicitly before anything moves.

That is why `safe_delete '*.log'` works while `rm *.log` is blocked: the

guard refuses opaque target sets, not cleanup work.

## If you run `rm` / destructive git anyway

A harness adapter may intercept the command before execution. Verdicts you

can receive:

| Verdict | Meaning | Your move |

|---|---|---|

| `PROCEED` | compensation already applied | continue; note the `txid` |

| `BLOCK_UNDETERMINABLE` | targets unresolvable (`$VAR`, `bash -c`, `find -delete`, `xargs`) | restate with explicit paths, or use `safe_delete` |

| `BLOCK_WILDCARD` | glob target set is opaque | use `safe_delete` with the glob |

| `BLOCK_OUT_OF_WORKSPACE` / `BLOCK_PROTECTED_PATH` | outside boundary, workspace root, or `.git` | do not retry; this is a hard boundary. Ask the human if it is truly needed |

| `BLOCK_RESTRICTED_MODE` | session is downgraded | only explicit single-file deletes are permitted; ask the human for anything more |

| `BLOCK_FORCE_PUSH` | remote history destruction | do not retry; escalate to the human |

A block is not an error to route around. Retrying the same operation in a

disguised form (`/bin/rm`, `python -c`, a script) is a violation of the

authorization pillar and is recorded in the audit log.

## RESTRICTED mode

After a human veto, the session runs with narrowed powers: explicit

single-file deletes inside the workspace still work (quarantined as usual);

recursive deletes, globs, and destructive git operations are refused. Only a

human can restore NORMAL. If a task genuinely requires more, say so plainly

and ask.

## Regenerable artifacts

Provably regenerable targets (git-ignored AND matching known artifact names

like `node_modules/`, `dist/`, `__pycache__/`) are allowed to be deleted

directly - no quarantine overhead. If git cannot prove a target is ignored,

it is treated as valuable and quarantined. Uncertainty increases restriction.

## Recovery

Every relocation writes `origin -> trash` pairs into

`.agent-trash/manifest.jsonl`. Restore is non-destructive: it refuses to

overwrite anything that now exists at the origin unless a human passes

`--force`. Git snapshots are stored as stashes named `agent-guard:<txid>`;

restore applies them and never drops them.

See `references/policy.md` for the complete rule table, verdict codes, and

manifest format.