diff --git a/harness/agents/impl/skills/gitea-issue-rework/SKILL.md b/harness/agents/impl/skills/gitea-issue-rework/SKILL.md
index cc85de1..fe609a1 100644
--- a/harness/agents/impl/skills/gitea-issue-rework/SKILL.md
+++ b/harness/agents/impl/skills/gitea-issue-rework/SKILL.md
@@ -1,7 +1,7 @@
 ---
 name: gitea-issue-rework
 description: Gitea PR 검토 지적을 계약에 맞춰 재작업하고 검증한다.
-version: 0.2.6
+version: 0.2.7
 author: genichin, Hermes Agent
 license: MIT
 platforms: [linux]
@@ -163,6 +163,10 @@ Reviewer가 rework head를 새 snapshot으로 다시 검토해야 하며 이 com
 
 ## 절차
 
+### Evidence manifest automation
+
+모든 baseline, 통합·precommit 및 readback 검증은 `scripts/evidence_manifest.py`와 `templates/evidence-plan.json`을 사용해 exact argv, exit code, 제한된 output, 허용된 환경, source revision, 파일 hash와 표준 JSON-pointer claims를 기록한다. 명령은 argv 배열만 허용하고 명시된 최소 child-environment allowlist에서 실행하며 output 상한을 넘는 process는 중단한다. Secret·header·private-key·password 값이나 plan 금지 패턴이 발견되면 redaction하지 않고 fail-closed한다. `run`은 baseline부터 최종 precommit까지 사용하고, `validate`는 plan과 manifest를 제외한 clean tree에서 PR claim 생성 전에 실행한다. `validate`는 source ancestry와 hash를 다시 검증하며 evidence-only descendant에서는 기록된 hash path가 source blob과 동일해야 한다. 수동으로 매칭되지 않는 PR 성공·검증 주장을 만들지 않는다. Comment publish 전 remote immutable artifact의 digest를 authoritative readback으로 비교한다.
+
 ### 1. Target, identity와 review 고정
 
 1. Exact target identifier를 검증하고 Issue lock을 획득한다.
diff --git a/harness/agents/impl/skills/gitea-issue-rework/scripts/evidence_manifest.py b/harness/agents/impl/skills/gitea-issue-rework/scripts/evidence_manifest.py
new file mode 100644
index 0000000..bb29395
--- /dev/null
+++ b/harness/agents/impl/skills/gitea-issue-rework/scripts/evidence_manifest.py
@@ -0,0 +1,319 @@
+#!/usr/bin/env python3
+"""Create and validate bounded, fail-closed evidence manifests."""
+from __future__ import annotations
+import argparse, datetime as dt, hashlib, json, math, os, platform, re, shutil, subprocess, sys, tempfile, time
+from pathlib import Path
+
+HEX40 = re.compile(r"^[0-9a-f]{40}$")
+HEX64 = re.compile(r"^[0-9a-f]{64}$")
+ID = re.compile(r"^[A-Za-z][A-Za-z0-9_.-]{0,127}$")
+ISO = re.compile(r"^\d{4}-\d\d-\d\dT\d\d:\d\d:\d\d(?:\.\d+)?Z$")
+SAFE_ENV = ("os", "kernel", "architecture", "python", "uv")
+SECRET_PATTERNS = [
+    re.compile(r"(?:cookie|set-cookie|authorization|proxy-authorization|x-api-key|api-key|x-auth-token)\s*:", re.I),
+    re.compile(r"\bbearer\s+[^\s]+", re.I),
+    re.compile(r"-----BEGIN [^-]*PRIVATE KEY-----", re.I),
+    re.compile(r"\b(?:password|passwd|secret|token)\s*[:=]", re.I),
+    re.compile(r"(?:^|\s)--?(?:password|passwd|secret|token)(?:\s+|=)\S+", re.I),
+]
+SAFE_CHILD_ENV = ("PATH", "HOME", "TMPDIR", "TMP", "TEMP", "LANG", "LC_ALL", "VIRTUAL_ENV", "UV_CACHE_DIR")
+SHELL_NAMES = {"sh", "bash", "dash", "zsh", "ksh", "fish", "cmd", "cmd.exe", "powershell", "powershell.exe", "pwsh", "pwsh.exe"}
+
+
+def die(msg):
+    print(str(msg), file=sys.stderr); return 2
+
+
+def git(root: Path, *args, check=True):
+    p = subprocess.run(["git", f"-c", f"safe.directory={root}", *args], cwd=root, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
+    if check and p.returncode: raise ValueError(p.stderr.decode("utf-8", "replace").strip() or "git failed")
+    return p
+
+
+def inside(root, value, *, allow_absolute=False):
+    if not isinstance(value, str) or not value or (Path(value).is_absolute() and not allow_absolute):
+        raise ValueError("paths must be nonempty repository-relative paths")
+    p = Path(value)
+    if ".." in p.parts: raise ValueError("path traversal is forbidden")
+    if p.parts and p.parts[0] == ".git": raise ValueError(".git paths are forbidden")
+    candidate = (root / p).resolve()
+    try: candidate.relative_to(root.resolve())
+    except ValueError: raise ValueError(f"path outside repository: {value}")
+    if (root / p).is_symlink() and not candidate.is_relative_to(root.resolve()): raise ValueError("symlink escapes repository")
+    return candidate, candidate.relative_to(root.resolve()).as_posix()
+
+
+def utc_now(): return dt.datetime.now(dt.timezone.utc).isoformat().replace("+00:00", "Z")
+def canonical(obj): return json.dumps(obj, sort_keys=True, separators=(",", ":"), ensure_ascii=False) + "\n"
+def text_bytes(b): return b.decode("utf-8", "replace")
+
+
+def scan_sensitive(text, secrets, forbidden):
+    for s in secrets:
+        if s and s in text: return "secret value"
+    for pattern in SECRET_PATTERNS + forbidden:
+        if pattern.search(text): return f"forbidden pattern: {pattern.pattern}"
+    return None
+
+
+def validate_regexes(values):
+    if not isinstance(values, list) or any(not isinstance(x, str) for x in values): raise ValueError("forbidden_regexes must be strings")
+    try: return [re.compile(x) for x in values]
+    except re.error as e: raise ValueError(f"invalid forbidden regex: {e}")
+
+
+def validate_argv(argv):
+    if not isinstance(argv,list) or not argv or any(not isinstance(x,str) for x in argv): raise ValueError("commands require nonempty argv string arrays")
+    executable=Path(argv[0]).name.lower()
+    if executable in SHELL_NAMES or any(x.lower() in ("shell", "shell=true") for x in argv): raise ValueError("shell commands are forbidden")
+
+
+def run_bounded(argv, root, env, max_bytes):
+    proc=subprocess.Popen(argv,cwd=root,stdout=subprocess.PIPE,stderr=subprocess.STDOUT,shell=False,env=env,bufsize=0)
+    stream=proc.stdout
+    if stream is None: raise ValueError("command output pipe unavailable")
+    data=bytearray()
+    try:
+        while True:
+            chunk=stream.read(min(65536,max_bytes-len(data)+1))
+            if not chunk: break
+            data.extend(chunk)
+            if len(data)>max_bytes:
+                proc.kill(); proc.wait()
+                raise ValueError("command output exceeds max_output_bytes")
+        return proc.wait(),bytes(data)
+    finally:
+        stream.close()
+        if proc.poll() is None:
+            proc.kill(); proc.wait()
+
+
+def status_paths(root):
+    raw = git(root, "status", "--porcelain=v1", "-z").stdout
+    out=[]
+    for entry in raw.split(b"\0"):
+        if not entry: continue
+        path = entry[3:] if len(entry) >= 3 else b""
+        if entry[:2] in (b"R ", b" R", b"R?"):
+            path = path.split(b"\0",1)[-1]
+        out.append(path.decode("utf-8", "surrogateescape"))
+    return out
+
+
+def check_clean(root, allowed):
+    bad = [p for p in status_paths(root) if p not in allowed]
+    if bad: raise ValueError("source tree is not clean: " + ", ".join(bad))
+
+
+def plan_checks(plan, root):
+    if not isinstance(plan, dict): raise ValueError("plan must be an object")
+    for key in ("hash_paths", "commands", "limitations", "secret_env_names", "forbidden_regexes", "command_env_allowlist"):
+        if key in plan and not isinstance(plan[key], list): raise ValueError(f"{key} must be a list")
+    names = plan.get("secret_env_names", [])
+    if any(not isinstance(n, str) for n in names): raise ValueError("secret_env_names must be strings")
+    secrets = [os.environ.get(n, "") for n in names if os.environ.get(n, "")]
+    forbidden = validate_regexes(plan.get("forbidden_regexes", []))
+    requested = plan.get("environment", list(SAFE_ENV[:4]))
+    if not isinstance(requested, list) or len(set(requested)) != len(requested) or any(x not in SAFE_ENV for x in requested): raise ValueError("invalid environment allowlist")
+    child_names=plan.get("command_env_allowlist",list(SAFE_CHILD_ENV))
+    if any(not isinstance(x,str) or not x or x not in SAFE_CHILD_ENV for x in child_names): raise ValueError("command_env_allowlist may contain only safe operational names")
+    child_names=list(dict.fromkeys([*child_names,*names]))
+    hash_paths=plan.get("hash_paths",[])
+    if not hash_paths or any(not isinstance(x,str) or not x for x in hash_paths): raise ValueError("hash_paths must be a nonempty string list")
+    command_plans=plan.get("commands",[])
+    if not command_plans: raise ValueError("commands must be nonempty")
+    seen=set(); command_ids=set(); commands=[]
+    for index,spec in enumerate(command_plans):
+        if not isinstance(spec, dict): raise ValueError("command must be object")
+        argv=spec.get("argv"); validate_argv(argv)
+        expected=spec.get("expected_exit_code",0); maxb=spec.get("max_output_bytes",65536)
+        if not isinstance(expected,int) or isinstance(expected,bool) or not isinstance(maxb,int) or maxb < 0: raise ValueError("invalid command bounds")
+        claims=spec.get("claims",[])
+        if not isinstance(claims,list) or not claims or any(not isinstance(x,str) or not ID.fullmatch(x) for x in claims): raise ValueError("claims must be a nonempty list of valid IDs")
+        if any(x in seen for x in claims): raise ValueError("duplicate claim ID")
+        command_id=spec.get("id",f"command-{index}"); purpose=spec.get("purpose","")
+        if not isinstance(command_id,str) or not ID.fullmatch(command_id) or command_id in command_ids: raise ValueError("invalid or duplicate command ID")
+        if not isinstance(purpose,str): raise ValueError("command purpose must be a string")
+        command_ids.add(command_id); seen.update(claims); commands.append((spec,argv,expected,maxb,claims))
+    return secrets, forbidden, requested, child_names, commands
+
+
+def tracked_blob(root, source, rel):
+    info=git(root,"ls-tree",source,"--",rel).stdout.decode("utf-8","replace").strip().split()
+    if len(info)<3 or info[1] not in ("blob",) or info[0] not in ("100644","100755"): raise ValueError(f"hash path is not tracked regular file: {rel}")
+    oid=info[2]; blob=git(root,"cat-file","blob",oid).stdout
+    return oid, hashlib.sha256(blob).hexdigest()
+
+
+def atomic_write(path, data):
+    if not path.parent.is_dir(): raise ValueError("output parent must already exist")
+    fd,tmp=tempfile.mkstemp(prefix="."+path.name+".", dir=path.parent)
+    try:
+        with os.fdopen(fd,"w",encoding="utf-8",newline="") as f:
+            f.write(data); f.flush(); os.fsync(f.fileno())
+        os.replace(tmp,path)
+        dir_fd = os.open(path.parent, os.O_RDONLY)
+        try: os.fsync(dir_fd)
+        finally: os.close(dir_fd)
+    finally:
+        try: os.unlink(tmp)
+        except FileNotFoundError: pass
+
+
+def run(args):
+    root=Path(args.repo_root).resolve(); plan_path,plan_rel=inside(root,args.plan); out,out_rel=inside(root,args.output)
+    if plan_path.suffix.lower()!=".json": raise ValueError("plan must be a JSON file")
+    if not out.parent.is_dir(): raise ValueError("output parent must already exist")
+    plan_bytes=plan_path.read_bytes(); plan_sha256=hashlib.sha256(plan_bytes).hexdigest()
+    plan=json.loads(plan_bytes.decode("utf-8")); secrets,forbidden,env_keys,child_names,commandspec=plan_checks(plan,root)
+    for _,argv,_,_,_ in commandspec:
+        violation=scan_sensitive("\n".join(argv),secrets,forbidden)
+        if violation: raise ValueError("sensitive command rejected before execution: "+violation)
+    check_clean(root,{plan_rel,out_rel})
+    source=git(root,"rev-parse","HEAD").stdout.decode().strip()
+    if not HEX40.fullmatch(source): raise ValueError("source revision is not full SHA")
+    files=[]
+    for item in plan.get("hash_paths",[]):
+        path,rel=inside(root,item)
+        if not path.is_file() or path.is_symlink(): raise ValueError(f"hash path is not a regular file: {item}")
+        oid,sha=tracked_blob(root,source,rel)
+        if hashlib.sha256(path.read_bytes()).hexdigest()!=sha: raise ValueError(f"source hash mismatch: {rel}")
+        files.append({"path":rel,"sha256":sha,"blob_oid":oid})
+    bound_paths={record["path"] for record in files}
+    if plan_rel==out_rel or plan_rel in bound_paths or out_rel in bound_paths: raise ValueError("plan/output paths must not overlap bound source paths")
+    child_env={name:os.environ[name] for name in child_names if name in os.environ}
+    result_commands=[]; claim_map={}
+    for i,(spec,argv,expected,maxb,claim_ids) in enumerate(commandspec):
+        started=utc_now(); t=time.monotonic(); returncode,raw=run_bounded(argv,root,child_env,maxb); finished=utc_now()
+        output=text_bytes(raw); violation=scan_sensitive("\n".join(argv)+"\n"+output,secrets,forbidden)
+        if violation: raise ValueError("sensitive evidence rejected: "+violation)
+        c={"id":spec.get("id",f"command-{i}"),"purpose":spec.get("purpose",""),"argv":argv,"exit_code":returncode,"expected_exit_code":expected,"max_output_bytes":maxb,"output":output,"started_at_utc":started,"finished_at_utc":finished,"duration_seconds":time.monotonic()-t,"claim_ids":claim_ids}
+        result_commands.append(c)
+        for cid in claim_ids:
+            claim_map[cid]=[
+                f"/commands/{i}/argv",
+                f"/commands/{i}/exit_code",
+                f"/commands/{i}/expected_exit_code",
+                f"/commands/{i}/output",
+                f"/commands/{i}/started_at_utc",
+                f"/commands/{i}/finished_at_utc",
+            ]
+    env={k:None for k in env_keys}
+    if "os" in env: env["os"]=platform.system()
+    if "kernel" in env: env["kernel"]=platform.release()
+    if "architecture" in env: env["architecture"]=platform.machine()
+    if "python" in env: env["python"]=platform.python_version()
+    if "uv" in env:
+        p=subprocess.run(["uv","--version"],stdout=subprocess.PIPE,stderr=subprocess.PIPE) if shutil.which("uv") else None
+        env["uv"]=text_bytes((p.stdout+p.stderr).strip()) if p and p.returncode==0 else None
+    # Commands are evidence producers, not trusted source mutators. Recheck the
+    # tree and every bound blob before persisting their claims.
+    check_clean(root,{plan_rel,out_rel})
+    if git(root,"rev-parse","HEAD").stdout.decode().strip()!=source: raise ValueError("HEAD changed during verification")
+    if not plan_path.is_file() or hashlib.sha256(plan_path.read_bytes()).hexdigest()!=plan_sha256: raise ValueError("plan changed during verification")
+    for record in files:
+        if hashlib.sha256((root/record["path"]).read_bytes()).hexdigest()!=record["sha256"]:
+            raise ValueError(f"source changed during verification: {record['path']}")
+    success=all(c["exit_code"]==c["expected_exit_code"] for c in result_commands)
+    manifest={"schema":"evidence-manifest/v1","source_revision":source,"source_tree_clean":True,"excluded_paths":[plan_rel,out_rel],"plan":{"path":plan_rel,"sha256":plan_sha256},"generated_at":utc_now(),"environment":env,"commands":result_commands,"files":files,"claims":claim_map,"redaction":{"checked":True,"violations":[]},"limitations":plan.get("limitations",[]),"result":"success" if success else "failure"}
+    atomic_write(out,canonical(manifest))
+    return 0 if success else 1
+
+
+def pointer(obj,ptr):
+    if not isinstance(ptr,str) or not ptr.startswith("/"): raise ValueError("invalid claim pointer")
+    cur=obj
+    for part in ptr[1:].split("/"):
+        part=part.replace("~1","/").replace("~0","~")
+        if isinstance(cur,list) and part.isdigit() and int(part)<len(cur): cur=cur[int(part)]
+        elif isinstance(cur,dict) and part in cur: cur=cur[part]
+        else: raise ValueError("claim pointer does not resolve")
+    return cur
+
+
+def validate(args):
+    root=Path(args.repo_root).resolve(); mp,mp_rel=inside(root,args.manifest); raw=mp.read_text(encoding="utf-8"); obj=json.loads(raw)
+    if canonical(obj)!=raw: raise ValueError("manifest is not canonical JSON")
+    required={"schema","source_revision","source_tree_clean","excluded_paths","plan","generated_at","commands","files","claims","redaction","environment","limitations","result"}
+    if set(obj)!=required or obj["schema"]!="evidence-manifest/v1" or obj["result"] not in ("success","failure"): raise ValueError("invalid manifest structure")
+    if obj["source_tree_clean"] is not True or not isinstance(obj["excluded_paths"],list): raise ValueError("source tree was not clean")
+    if not isinstance(obj["generated_at"],str) or not ISO.fullmatch(obj["generated_at"]): raise ValueError("invalid generated timestamp")
+    if len(obj["excluded_paths"])!=2 or len(set(obj["excluded_paths"]))!=2 or any(not isinstance(x,str) for x in obj["excluded_paths"]) or obj["excluded_paths"][1]!=mp_rel: raise ValueError("invalid excluded paths")
+    for excluded in obj["excluded_paths"]: inside(root,excluded)
+    if not isinstance(obj["plan"],dict) or set(obj["plan"])!={"path","sha256"} or obj["plan"]["path"]!=obj["excluded_paths"][0] or not isinstance(obj["plan"]["sha256"],str) or not HEX64.fullmatch(obj["plan"]["sha256"]): raise ValueError("invalid plan binding")
+    plan_path,plan_rel=inside(root,obj["plan"]["path"])
+    if plan_path.suffix.lower()!=".json" or not plan_path.is_file() or plan_path.is_symlink(): raise ValueError("bound plan is unavailable")
+    plan_bytes=plan_path.read_bytes()
+    if hashlib.sha256(plan_bytes).hexdigest()!=obj["plan"]["sha256"]: raise ValueError("plan hash mismatch")
+    plan=json.loads(plan_bytes.decode("utf-8")); plan_secrets,plan_forbidden,requested_env,_,planned_commands=plan_checks(plan,root)
+    check_clean(root,set(obj["excluded_paths"]))
+    src=obj["source_revision"]
+    if not isinstance(src,str) or not HEX40.fullmatch(src): raise ValueError("invalid source revision")
+    if git(root,"cat-file","-e",src,check=False).returncode: raise ValueError("missing source revision")
+    head=git(root,"rev-parse","HEAD").stdout.decode().strip()
+    if git(root,"merge-base","--is-ancestor",src,head,check=False).returncode: raise ValueError("source is not ancestor")
+    if not isinstance(obj["environment"],dict) or not obj["environment"] or any(not isinstance(k,str) or k not in SAFE_ENV or not (isinstance(v,str) or v is None) for k,v in obj["environment"].items()): raise ValueError("invalid environment")
+    if set(obj["environment"])!=set(requested_env): raise ValueError("environment does not match plan")
+    if not isinstance(obj["limitations"],list) or any(not isinstance(x,str) for x in obj["limitations"]) or obj["limitations"]!=plan.get("limitations",[]): raise ValueError("invalid limitations")
+    if obj["redaction"]!={"checked":True,"violations":[]}: raise ValueError("redaction failed")
+    if not isinstance(obj["claims"],dict) or any(
+        not isinstance(k,str) or not ID.fullmatch(k) or not isinstance(v,list)
+        or len(v)!=6 or any(not isinstance(ptr,str) for ptr in v)
+        for k,v in obj["claims"].items()
+    ): raise ValueError("invalid claims")
+    if not isinstance(obj["files"],list) or not obj["files"]: raise ValueError("files must be nonempty")
+    planned_paths=[inside(root,item)[1] for item in plan.get("hash_paths",[])]
+    if [f.get("path") if isinstance(f,dict) else None for f in obj["files"]]!=planned_paths: raise ValueError("file records do not match plan")
+    claim_seen=set()
+    for f in obj["files"]:
+        if not isinstance(f,dict) or set(f)!={"path","sha256","blob_oid"}: raise ValueError("invalid file record")
+        _,rel=inside(root,f["path"]); path=root/rel
+        if not HEX64.fullmatch(f["sha256"]) or not HEX40.fullmatch(f["blob_oid"]): raise ValueError("invalid file identity")
+        oid,sha=tracked_blob(root,src,rel)
+        if oid!=f["blob_oid"] or sha!=f["sha256"] or hashlib.sha256(path.read_bytes()).hexdigest()!=sha: raise ValueError("file hash mismatch")
+    if any(f["path"] in set(obj["excluded_paths"]) for f in obj["files"]): raise ValueError("evidence paths overlap bound source paths")
+    if not isinstance(obj["commands"],list) or not obj["commands"] or len(obj["commands"])!=len(planned_commands): raise ValueError("commands do not match plan")
+    command_ids=set()
+    for i,c in enumerate(obj["commands"]):
+        spec,planned_argv,planned_exit,planned_max,planned_claims=planned_commands[i]
+        command_keys={"id","purpose","argv","exit_code","expected_exit_code","max_output_bytes","output","started_at_utc","finished_at_utc","duration_seconds","claim_ids"}
+        if not isinstance(c,dict) or set(c)!=command_keys: raise ValueError("invalid command structure")
+        if not isinstance(c["id"],str) or not ID.fullmatch(c["id"]) or c["id"] in command_ids or not isinstance(c["purpose"],str): raise ValueError("invalid command identity")
+        command_ids.add(c["id"])
+        validate_argv(c["argv"])
+        if not isinstance(c["output"],str): raise ValueError("invalid command types")
+        if any(not isinstance(c[k],int) or isinstance(c[k],bool) for k in ("exit_code","expected_exit_code","max_output_bytes")) or c["max_output_bytes"]<0 or len(c["output"].encode())>c["max_output_bytes"]: raise ValueError("invalid command bounds")
+        if not isinstance(c["duration_seconds"],(int,float)) or isinstance(c["duration_seconds"],bool) or not math.isfinite(c["duration_seconds"]) or c["duration_seconds"]<0: raise ValueError("invalid command duration")
+        if not isinstance(c["claim_ids"],list) or any(not isinstance(x,str) or not ID.fullmatch(x) for x in c["claim_ids"]): raise ValueError("invalid command fields")
+        if c["id"]!=spec.get("id",f"command-{i}") or c["purpose"]!=spec.get("purpose","") or c["argv"]!=planned_argv or c["expected_exit_code"]!=planned_exit or c["max_output_bytes"]!=planned_max or c["claim_ids"]!=planned_claims: raise ValueError("command record does not match plan")
+        if not ISO.fullmatch(c["started_at_utc"]) or not ISO.fullmatch(c["finished_at_utc"]): raise ValueError("invalid timestamp")
+        for cid in c["claim_ids"]:
+            if cid in claim_seen or cid not in obj["claims"]: raise ValueError("duplicate or missing claim")
+            claim_seen.add(cid)
+            pointers=obj["claims"][cid]
+            expected_pointers=[
+                f"/commands/{i}/argv",
+                f"/commands/{i}/exit_code",
+                f"/commands/{i}/expected_exit_code",
+                f"/commands/{i}/output",
+                f"/commands/{i}/started_at_utc",
+                f"/commands/{i}/finished_at_utc",
+            ]
+            if pointers!=expected_pointers: raise ValueError("claim pointers are not canonical")
+            for ptr in pointers: pointer(obj,ptr)
+        violation=scan_sensitive("\n".join(c["argv"])+"\n"+c["output"],plan_secrets,plan_forbidden)
+        if violation: raise ValueError("redaction violation")
+    if claim_seen != set(obj["claims"]): raise ValueError("unreferenced claim")
+    expected="success" if all(c["exit_code"]==c["expected_exit_code"] for c in obj["commands"]) else "failure"
+    if obj["result"]!=expected: raise ValueError("result inconsistent")
+    return 0
+
+
+def main():
+    p=argparse.ArgumentParser(); sub=p.add_subparsers(dest="op",required=True)
+    r=sub.add_parser("run"); r.add_argument("--repo-root",required=True); r.add_argument("--plan",required=True); r.add_argument("--output",required=True)
+    v=sub.add_parser("validate"); v.add_argument("--repo-root",required=True); v.add_argument("--manifest",required=True)
+    a=p.parse_args()
+    try: return run(a) if a.op=="run" else validate(a)
+    except (OSError, ValueError, TypeError, KeyError, AttributeError, UnicodeError, json.JSONDecodeError, re.error, subprocess.SubprocessError) as e: return die(e)
+if __name__=="__main__": sys.exit(main())
diff --git a/harness/agents/impl/skills/gitea-issue-rework/templates/evidence-plan.json b/harness/agents/impl/skills/gitea-issue-rework/templates/evidence-plan.json
new file mode 100644
index 0000000..790d763
--- /dev/null
+++ b/harness/agents/impl/skills/gitea-issue-rework/templates/evidence-plan.json
@@ -0,0 +1,18 @@
+{
+  "environment": ["os", "kernel", "architecture", "python"],
+  "command_env_allowlist": ["PATH", "HOME", "TMPDIR", "LANG", "VIRTUAL_ENV", "UV_CACHE_DIR"],
+  "secret_env_names": ["PROJECT_TEST_TOKEN"],
+  "forbidden_regexes": [],
+  "hash_paths": ["src/project.py", "tests/test_project.py"],
+  "commands": [
+    {
+      "id": "focused-tests",
+      "purpose": "Run the focused verification",
+      "argv": ["python3", "-m", "unittest", "tests.test_project", "-v"],
+      "expected_exit_code": 0,
+      "max_output_bytes": 65536,
+      "claims": ["focused-result"]
+    }
+  ],
+  "limitations": []
+}
diff --git a/harness/tests/impl/test_evidence_manifest.py b/harness/tests/impl/test_evidence_manifest.py
new file mode 100644
index 0000000..5b5cb3e
--- /dev/null
+++ b/harness/tests/impl/test_evidence_manifest.py
@@ -0,0 +1,217 @@
+import hashlib
+import json
+import os
+import subprocess
+import sys
+import tempfile
+import unittest
+from pathlib import Path
+
+SCRIPT = Path(__file__).resolve().parents[2] / "agents/impl/skills/gitea-issue-rework/scripts/evidence_manifest.py"
+
+
+def run_git(cwd, *args):
+    return subprocess.run(["git", *args], cwd=cwd, text=True, capture_output=True, check=True)
+
+
+class EvidenceManifestTests(unittest.TestCase):
+    def setUp(self):
+        self.tmp = tempfile.TemporaryDirectory()
+        self.repo = Path(self.tmp.name) / "repo"
+        self.repo.mkdir()
+        run_git(self.repo, "init", "-q")
+        run_git(self.repo, "config", "user.email", "test@example.invalid")
+        run_git(self.repo, "config", "user.name", "Test")
+        (self.repo / "input.txt").write_text("hello\n")
+        run_git(self.repo, "add", "input.txt")
+        run_git(self.repo, "commit", "-qm", "initial")
+
+    def tearDown(self): self.tmp.cleanup()
+
+    def invoke(self, *args, env=None):
+        return subprocess.run([sys.executable, str(SCRIPT), *map(str, args)], cwd=self.repo, text=True, capture_output=True, env=env)
+
+    def plan(self, commands=None, **extra):
+        if commands is None:
+            commands = [{"id": "check", "purpose": "show input", "argv": ["git", "show", "HEAD:input.txt"], "expected_exit_code": 0, "claims": ["output"]}]
+        data = {"hash_paths": ["input.txt"], "commands": commands}
+        data.update(extra); p = self.repo / "plan.json"; p.write_text(json.dumps(data)); return p
+
+    def run_manifest(self, plan=None):
+        if plan is None: plan = self.plan()
+        out = self.repo / "manifest.json"; result = self.invoke("run", "--repo-root", self.repo, "--plan", plan.name, "--output", "manifest.json"); return result, out
+
+    def test_run_captures_success_claim_and_hash(self):
+        result, out = self.run_manifest(); self.assertEqual(result.returncode, 0, result.stderr)
+        manifest = json.loads(out.read_text())
+        self.assertEqual(manifest["commands"][0]["argv"], ["git", "show", "HEAD:input.txt"])
+        self.assertEqual(manifest["commands"][0]["exit_code"], 0)
+        self.assertEqual(manifest["commands"][0]["claim_ids"], ["output"])
+        self.assertEqual(
+            manifest["claims"]["output"],
+            [
+                "/commands/0/argv",
+                "/commands/0/exit_code",
+                "/commands/0/expected_exit_code",
+                "/commands/0/output",
+                "/commands/0/started_at_utc",
+                "/commands/0/finished_at_utc",
+            ],
+        )
+        self.assertEqual(manifest["files"][0]["sha256"], hashlib.sha256(b"hello\n").hexdigest())
+        self.assertEqual(self.invoke("validate", "--repo-root", self.repo, "--manifest", "manifest.json").returncode, 0)
+
+    def test_unexpected_exit_is_recorded_and_nonzero(self):
+        result, out = self.run_manifest(self.plan([{"argv": [sys.executable, "-c", "print('no'); raise SystemExit(3)"], "expected_exit_code": 0, "claims": ["result"]}]))
+        self.assertNotEqual(result.returncode, 0); self.assertEqual(json.loads(out.read_text())["commands"][0]["exit_code"], 3)
+
+    def test_failed_run_preserves_prior_manifest(self):
+        result, out = self.run_manifest()
+        self.assertEqual(result.returncode, 0, result.stderr)
+        prior = out.read_bytes()
+        bad = self.plan([{"argv": [sys.executable, "-c", "print('Authorization: Bearer ***')"]}])
+        result = self.invoke("run", "--repo-root", self.repo, "--plan", bad, "--output", out)
+        self.assertNotEqual(result.returncode, 0)
+        self.assertEqual(out.read_bytes(), prior)
+
+    def test_command_environment_is_allowlisted_and_output_is_bounded(self):
+        env = os.environ.copy()
+        env["UNLISTED_SECRET"] = "must-not-reach-child"
+        command = [{
+            "argv": [sys.executable, "-c", "import os; print(os.getenv('UNLISTED_SECRET'))"],
+            "claims": ["environment"],
+        }]
+        plan = self.plan(command)
+        out = self.repo / "manifest.json"
+        result = self.invoke("run", "--repo-root", self.repo, "--plan", plan.name, "--output", out.name, env=env)
+        self.assertEqual(result.returncode, 0, result.stderr)
+        self.assertEqual(json.loads(out.read_text())["commands"][0]["output"], "None\n")
+
+        env["LEAK"] = "SECRET_VALUE"
+        plan = self.plan(command, command_env_allowlist=["LEAK"])
+        result = self.invoke("run", "--repo-root", self.repo, "--plan", plan.name, "--output", out.name, env=env)
+        self.assertNotEqual(result.returncode, 0)
+        self.assertNotIn("SECRET_VALUE", out.read_text())
+
+        prior = out.read_bytes()
+        plan = self.plan([{
+            "argv": [sys.executable, "-c", "print('x' * 100000)"],
+            "max_output_bytes": 16,
+            "claims": ["bounded"],
+        }])
+        result = self.invoke("run", "--repo-root", self.repo, "--plan", plan.name, "--output", out.name)
+        self.assertNotEqual(result.returncode, 0)
+        self.assertEqual(out.read_bytes(), prior)
+
+    def test_output_collision_and_password_argument_are_rejected(self):
+        plan = self.plan()
+        result = self.invoke("run", "--repo-root", self.repo, "--plan", plan.name, "--output", "input.txt")
+        self.assertNotEqual(result.returncode, 0)
+        self.assertEqual((self.repo / "input.txt").read_text(), "hello\n")
+
+        result, out = self.run_manifest(self.plan([{
+            "argv": [sys.executable, "-c", "from pathlib import Path; Path('ran.txt').write_text('ran')", "--password", "visible-value"],
+            "claims": ["password-leak"],
+        }]))
+        self.assertNotEqual(result.returncode, 0)
+        self.assertFalse(out.exists())
+        self.assertFalse((self.repo / "ran.txt").exists())
+
+    def test_argv_only_and_traversal_rejected(self):
+        result, out = self.run_manifest(self.plan([{"argv": ["echo", "shell=true"]}]))
+        self.assertNotEqual(result.returncode, 0); self.assertFalse(out.exists())
+        result, out = self.run_manifest(self.plan(hash_paths=["../outside"]))
+        self.assertNotEqual(result.returncode, 0); self.assertFalse(out.exists())
+        result, out = self.run_manifest(self.plan(hash_paths=["nested/../input.txt"]))
+        self.assertNotEqual(result.returncode, 0); self.assertFalse(out.exists())
+
+    def test_empty_plan_and_malformed_command_record_are_rejected(self):
+        result, out = self.run_manifest(self.plan([]))
+        self.assertNotEqual(result.returncode, 0)
+        self.assertFalse(out.exists())
+
+        result, out = self.run_manifest()
+        self.assertEqual(result.returncode, 0, result.stderr)
+        data = json.loads(out.read_text())
+        data["commands"][0]["duration_seconds"] = -1
+        out.write_text(json.dumps(data, sort_keys=True, separators=(",", ":")) + "\n")
+        self.assertNotEqual(self.invoke("validate", "--repo-root", self.repo, "--manifest", out.name).returncode, 0)
+
+    def test_secret_or_header_leak_fails_closed(self):
+        env = os.environ.copy(); env["TEST_SECRET"] = "never-write-this"
+        p = self.plan([{"argv": [sys.executable, "-c", "import os; print(os.environ['TEST_SECRET'])"]}], secret_env_names=["TEST_SECRET"])
+        result = self.invoke("run", "--repo-root", self.repo, "--plan", "plan.json", "--output", "manifest.json", env=env)
+        self.assertNotEqual(result.returncode, 0); self.assertFalse((self.repo / "manifest.json").exists())
+
+    def test_source_drift_hash_mismatch_rejected(self):
+        result, out = self.run_manifest(); self.assertEqual(result.returncode, 0)
+        (self.repo / "input.txt").write_text("changed\n")
+        self.assertNotEqual(self.invoke("validate", "--repo-root", self.repo, "--manifest", out.name).returncode, 0)
+
+    def test_command_source_mutation_is_rejected_before_manifest_write(self):
+        command = [{
+            "argv": [sys.executable, "-c", "from pathlib import Path; Path('input.txt').write_text('changed\\n')"],
+            "claims": ["mutation"],
+        }]
+        result, out = self.run_manifest(self.plan(command))
+        self.assertNotEqual(result.returncode, 0)
+        self.assertFalse(out.exists())
+
+    def test_command_commit_drift_is_rejected_before_manifest_write(self):
+        script = (
+            "from pathlib import Path; import subprocess; "
+            "Path('other.txt').write_text('committed\\n'); "
+            "subprocess.run(['git','add','other.txt'],check=True); "
+            "subprocess.run(['git','commit','-qm','drift'],check=True)"
+        )
+        result, out = self.run_manifest(self.plan([{
+            "argv": [sys.executable, "-c", script],
+            "claims": ["commit-drift"],
+        }]))
+        self.assertNotEqual(result.returncode, 0)
+        self.assertFalse(out.exists())
+
+    def test_binary_source_hash_is_byte_accurate(self):
+        payload = b"\xff\x00\x80binary\n"
+        (self.repo / "binary.bin").write_bytes(payload)
+        run_git(self.repo, "add", "binary.bin")
+        run_git(self.repo, "commit", "-qm", "binary")
+        result, out = self.run_manifest(self.plan(hash_paths=["binary.bin"]))
+        self.assertEqual(result.returncode, 0, result.stderr)
+        self.assertEqual(json.loads(out.read_text())["files"][0]["sha256"], hashlib.sha256(payload).hexdigest())
+
+    def test_invalid_claim_pointer_rejected(self):
+        result, out = self.run_manifest(self.plan([{"argv": ["true"], "claims": ["bad"]}]))
+        self.assertEqual(result.returncode, 0)
+        data=json.loads(out.read_text()); data["claims"]["bad"]="/missing"; out.write_text(json.dumps(data, sort_keys=True, separators=(",", ":"))+"\n")
+        self.assertNotEqual(self.invoke("validate", "--repo-root", self.repo, "--manifest", out.name).returncode, 0)
+
+    def test_validate_rejects_forged_plan_exclusion(self):
+        result, out = self.run_manifest()
+        self.assertEqual(result.returncode, 0, result.stderr)
+        (self.repo / "plan.json").unlink()
+        (self.repo / "dirty.txt").write_text("dirty\n")
+        data = json.loads(out.read_text())
+        data["excluded_paths"][0] = "dirty.txt"
+        out.write_text(json.dumps(data, sort_keys=True, separators=(",", ":")) + "\n")
+        self.assertNotEqual(self.invoke("validate", "--repo-root", self.repo, "--manifest", out.name).returncode, 0)
+
+    def test_validate_rejects_empty_or_shell_argv(self):
+        result, out = self.run_manifest()
+        self.assertEqual(result.returncode, 0, result.stderr)
+        data = json.loads(out.read_text())
+        data["commands"][0]["argv"] = []
+        out.write_text(json.dumps(data, sort_keys=True, separators=(",", ":")) + "\n")
+        self.assertNotEqual(self.invoke("validate", "--repo-root", self.repo, "--manifest", out.name).returncode, 0)
+
+        data["commands"][0]["argv"] = ["bash", "-c", "true"]
+        out.write_text(json.dumps(data, sort_keys=True, separators=(",", ":")) + "\n")
+        self.assertNotEqual(self.invoke("validate", "--repo-root", self.repo, "--manifest", out.name).returncode, 0)
+
+    def test_evidence_only_descendant_validates(self):
+        result, out = self.run_manifest(); self.assertEqual(result.returncode, 0)
+        (self.repo / "evidence.txt").write_text("evidence\n"); run_git(self.repo, "add", "evidence.txt"); run_git(self.repo, "commit", "-qm", "evidence")
+        self.assertEqual(self.invoke("validate", "--repo-root", self.repo, "--manifest", "manifest.json").returncode, 0)
+
+
+if __name__ == "__main__": unittest.main()