Post-Implementation Audit

Post-Implementation Audit

Read-only audit of finished code changes that checks requirements, correctness, regressions, verification and commit scope.

Skill files1 file
SKILL.md197 lines
1
---
2
name: post-implementation-audit
3
description: Read-only audit of completed code changes. Use after implementation or remediation to verify requirements, correctness, regressions, relevant verification, and commit scope. Do not use to implement or fix changes.
4
---
5
6
# Post-Implementation Audit
7
8
Independently audit completed code changes using available evidence.
9
10
This workflow is read-only. Find meaningful problems when they exist; do not manufacture findings.
11
12
## Scope and boundary
13
14
Follow the current task, applicable repository instructions such as `AGENTS.md`, and the actual implementation context.
15
16
Do not broaden the audit merely because additional review is possible.
17
18
Do not implement, remediate, refactor, stage, commit, or intentionally modify repository state while this workflow is active.
19
20
If a required verification step would intentionally modify tracked files, do not run it during the audit. Report it as a verification limitation and return that work to an implementation/remediation phase.
21
22
Unexpected side effects from otherwise appropriate verification commands must be reported, not reverted or cleaned up.
23
24
Do not create an audit file unless explicitly requested.
25
26
## 1. Establish target and baseline
27
28
Determine what change is actually being audited before judging it.
29
30
When Git is available, prefer the baseline in this order:
31
32
1. explicit baseline or commit range from the task;
33
2. a known implementation start point supported by context;
34
3. clearly attributable staged or working-tree changes.
35
36
Never select an arbitrary number of recent commits as the baseline.
37
38
Distinguish implementation changes from pre-existing or unrelated repository changes.
39
40
If the boundary cannot be established reliably, state the limitation and audit only what can be attributed with reasonable confidence.
41
42
Do not invent missing requirements, acceptance criteria, history, or implementation boundaries.
43
44
## 2. Verify requirements, correctness, and regressions
45
46
Trace available requirements and acceptance criteria to the implementation.
47
48
Check for meaningful issues such as:
49
50
- missing or partial behavior;
51
- incorrect requirement interpretation;
52
- regressions or unintended behavior changes;
53
- scope creep or unrelated modifications;
54
- incorrect logic or state transitions;
55
- relevant error or failure paths;
56
- plausible boundary, lifecycle, async, concurrency, persistence, caching, or integration problems.
57
58
Inspect enough surrounding code to understand the changed behavior.
59
60
Only investigate risk areas that are plausible for the implementation.
61
62
Review security, performance, data integrity, or deployment concerns only when the change makes them relevant.
63
64
Do not report subjective style preferences or speculative possibilities as defects.
65
66
A maintainability concern is a finding only when it creates a concrete correctness, reviewability, change-safety, or long-term engineering risk.
67
68
## 3. Verify evidence
69
70
Run the smallest relevant set of non-mutating verification commands needed for confidence.
71
72
Expand verification when scope or risk warrants it.
73
74
Never claim that a command, test, path, or behavior was verified unless it actually was.
75
76
If important verification cannot be completed, record:
77
78
- what was not verified;
79
- why;
80
- what confidence is lost.
81
82
A verification limitation is not automatically a finding.
83
84
Treat it as a finding only when the missing verification itself violates an explicit requirement or represents a concrete defect.
85
86
## 4. Review commits when applicable
87
88
When commits are part of the audited implementation, verify that each represents one coherent concern and is independently understandable, reviewable, and reasonably revertible.
89
90
Report material problems such as:
91
92
- unrelated concerns mixed together;
93
- hidden scope expansion;
94
- accidental unrelated changes;
95
- misleading commit boundaries;
96
- excessive size that materially harms reviewability or rollback safety.
97
98
Do not require commits when none were authorized or expected.
99
100
## 5. Complete the full audit
101
102
Do not stop at the first issue.
103
104
Complete the full in-scope review and collect every meaningful finding supported by evidence.
105
106
Each finding must be backed by code, diff, test output, command output, reproducible behavior, or a credible demonstrated failure path.
107
108
Distinguish fact from inference.
109
110
After completing the audit, report the result and stop. Do not remediate findings.
111
112
## Severity
113
114
Use severity only for actual findings:
115
116
**Critical** — catastrophic failure, severe security compromise, irreversible data loss/corruption, or fundamentally unusable core behavior.
117
118
**High** — major incorrect behavior, serious regression, significant security/reliability failure, or failure of an important requirement.
119
120
**Medium** — real actionable defect with bounded impact.
121
122
**Low** — minor but legitimate defect with limited concrete impact.
123
124
Do not use `Low` for optional polish or subjective preference.
125
126
## Result
127
128
Use exactly one primary result:
129
130
### CLEAR
131
132
Use when:
133
134
- no meaningful finding remains;
135
- intended behavior is sufficiently established;
136
- relevant verification completed successfully;
137
- no material unexplained verification gap remains;
138
- no material scope contamination exists.
139
140
### FINDINGS
141
142
Use when one or more meaningful implementation findings exist.
143
144
Verification limitations may be reported alongside `FINDINGS`.
145
146
### INCOMPLETE
147
148
Use when no meaningful implementation defect has been established, but missing context or important verification prevents a reliable `CLEAR`.
149
150
Do not treat absence of discovered defects as proof of correctness.
151
152
## Re-audit
153
154
When previous findings are available, preserve their identifiers and mark each:
155
156
- `RESOLVED`
157
- `UNRESOLVED`
158
- `NOT VERIFIED`
159
160
Verify the underlying issue, not only its visible symptom, and check whether remediation introduced regressions.
161
162
Then perform a fresh audit of the affected scope.
163
164
Do not invent prior finding IDs when they are unavailable.
165
166
## Output
167
168
Start with:
169
170
**Result:** `CLEAR` / `FINDINGS` / `INCOMPLETE`
171
172
Briefly state:
173
174
- scope audited;
175
- baseline used;
176
- important evidence inspected;
177
- verification commands actually executed;
178
- material limitations.
179
180
For each new finding include:
181
182
**ID:** `AUDIT-001`
183
**Severity:** Critical / High / Medium / Low
184
**Evidence:** concrete supporting evidence
185
**Impact:** concrete failure or risk
186
**Recommended remediation:** smallest appropriate correction
187
**Verification:** how a re-audit can prove resolution
188
189
For re-audited findings also include:
190
191
**Status:** RESOLVED / UNRESOLVED / NOT VERIFIED
192
193
For `INCOMPLETE`, state what evidence is missing.
194
195
For `CLEAR`, state that no meaningful findings remain and summarize the evidence supporting that conclusion.
196
197
Do not invent owners, deadlines, metrics, findings, or recommendations merely to make the report appear more comprehensive.
@OpenSourceBuffGuy

I find openly licensed prompts and post them here.

0 followers · 2,873 prompts

More by @OpenSourceBuffGuy

See all
Ethereum Developer prompt image

Ethereum Developer

Asks the AI to act as an experienced Ethereum developer who writes a Solidity smart contract for an on-chain messenger with public reads and owner-only writes.

act-as, crypto, coding

1 view
Linux Terminal prompt image

Linux Terminal

Makes the AI act as a Linux terminal that replies to each typed command with only its output inside one code block.

simulator, linux, coding

5 views
English Translator and Improver prompt image

English Translator and Improver

Turns the AI into an English translator and editor that detects the input language and returns only an improved, more literary English version.

act-as, translation, rewriting, proofreading

2 views

Reviews

Loading reviews…

Discussion

Loading comments…