retro

Name: retro
Author: liwala

84stars12forks

How to install retro?

liwala/sheal/retro

$npx -y skills add liwala/sheal --skill retro

Installs into the current project.

Summary

Run a deep session retrospective analyzing the most recent AI coding session. Uses Entire.io session data + static analysis, then the agent reviews its own work to extract learnings and generate rules.

Files · 1

View on GitHub

SKILL.md

1	# Session Retrospective Skill
2
3	You are performing a deep retrospective analysis of a completed AI coding session. This is the self-healing loop: you're reviewing what happened and generating actionable rules to improve future sessions.
4
5	## Step 1: Gather Data
6
7	Run the static analysis first. If a checkpoint ID was provided as an argument, use it. Otherwise, use the latest checkpoint.
8
9	```bash
10	sheal retro --format json [-c CHECKPOINT_ID]
11	```
12
13	Also load the session details to see the Entire.io summary:
14
15	```bash
16	sheal sessions --format json [-c CHECKPOINT_ID]
17	```
18
19	## Step 2: Deep Analysis
20
21	With the static analysis and session data loaded, analyze the session deeply. Consider:
22
23	### Failure Patterns
24	- For each failure loop or Bash failure: why did it happen? Was it a wrong approach, missing context, environment issue, or normal iteration?
25	- Could a pre-check have prevented it?
26	- Was the agent stuck or making progress?
27
28	### Effort Quality
29	- Was the file churn productive iteration or wasted effort?
30	- Were the right tools used? (e.g., using Bash when Read/Grep would have been better)
31	- Was research done before implementation, or was the approach trial-and-error?
32
33	### Missing Context
34	- What information was missing at the start that caused problems later?
35	- What should have been in CLAUDE.md, .cursorrules, or project documentation?
36	- Were there assumptions that turned out wrong?
37
38	### Workflow Improvements
39	- What would make the next session smoother?
40	- Are there recurring patterns that should become rules?
41	- Should any health checks be added to `sheal check`?
42
43	## Step 3: Generate Output
44
45	Present your analysis as a structured retrospective report:
46
47	### Report Format
48
49	```
50	## Session Retrospective: [checkpoint-id]
51
52	### Summary
53	[1-2 sentence summary of what happened and how it went]
54
55	### Health Score: [X/100]
56	[Explain why this score, what went well, what didn't]
57
58	### Key Findings
59
60	#### What Went Well
61	- [thing 1]
62	- [thing 2]
63
64	#### What Could Be Improved
65	- [issue 1]: [why it happened] → [what to do differently]
66	- [issue 2]: [why it happened] → [what to do differently]
67
68	### Suggested Rules
69	[List specific rules that should be added to agent config files]
70	```
71
72	## Step 4: Apply Learnings (with user confirmation)
73
74	After presenting the report, ask the user if they want to apply any of the suggested rules. If yes:
75
76	1. For Claude Code: append rules to `CLAUDE.md` in the project root
77	2. For Cursor: append to `.cursorrules`
78	3. For Gemini: append to `.gemini/GEMINI.md`
79	4. For general: append to `AGENTS.md`
80
81	Format rules as clear, actionable instructions. For example:
82
83	```markdown
84	## Learnings from session [date]
85
86	- Before writing parsers for external data formats, inspect real data samples first (`git show branch:path/to/file \| head -20`). Don't assume formats from documentation alone.
87	- When a CLI tool fails with "connection refused", check if its required server process is running before retrying the same command.
88	- Run `sheal check` at session start to catch environment issues early.
89	```
90
91	## Important Notes
92
93	- Be honest about what went well — not every session has problems
94	- Focus on actionable, specific learnings, not generic advice
95	- Rules should be phrased as instructions to a future AI agent
96	- Don't generate rules for one-off issues that won't recur
97	- Consider whether the issue is project-specific or universal

People who liked this also viewed

6 picks

Type

azure-diagnosticsDebug Azure production issues on Azure using AppLens, Azure Monitor, resource health, and safe triage. WHEN: debug production issues, troubleshoot app service, app service high CPU, app service deploySkillsJul 2026472k1.3k
systematic-debuggingUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesSkillsJul 2026197k259k
safe-debugRigor Debug / Rigor Audit skill for deep learning research work. Use when the user pastes a traceback, terminal error, CUDA OOM, checkpoint load failure, shape mismatch, NaN loss symptom, or training SkillsMay 2026161k507
explore-runRigor Improve / Rigor Explore run leaf skill for bounded exploratory evidence in deep learning research repositories. Use when the researcher explicitly authorizes exploratory runs such as small-subseSkillsMay 2026161k507
minimal-run-and-auditRigor Run skill for README-first deep learning repo reproduction. Use when the task is specifically to capture or normalize evidence from the selected smoke test or documented inference or evaluation SkillsMay 2026161k507
insforge-debugUse when diagnosing problems in an InsForge project — reactive failures (SDK error object, HTTP 4xx/5xx, gateway timeout 502/503/504, edge function failure or timeout, login/OAuth/auth errors, RLS denSkillsJul 202618k31

retro

84stars12forks

View on GitHub

How to install retro?

liwala/sheal/retro

$npx -y skills add liwala/sheal --skill retro

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Session Retrospective Skill
2
3	You are performing a deep retrospective analysis of a completed AI coding session. This is the self-healing loop: you're reviewing what happened and generating actionable rules to improve future sessions.
4
5	## Step 1: Gather Data
6
7	Run the static analysis first. If a checkpoint ID was provided as an argument, use it. Otherwise, use the latest checkpoint.
8
9	```bash
10	sheal retro --format json [-c CHECKPOINT_ID]
11	```
12
13	Also load the session details to see the Entire.io summary:
14
15	```bash
16	sheal sessions --format json [-c CHECKPOINT_ID]
17	```
18
19	## Step 2: Deep Analysis
20
21	With the static analysis and session data loaded, analyze the session deeply. Consider:
22
23	### Failure Patterns
24	- For each failure loop or Bash failure: why did it happen? Was it a wrong approach, missing context, environment issue, or normal iteration?
25	- Could a pre-check have prevented it?
26	- Was the agent stuck or making progress?
27
28	### Effort Quality
29	- Was the file churn productive iteration or wasted effort?
30	- Were the right tools used? (e.g., using Bash when Read/Grep would have been better)
31	- Was research done before implementation, or was the approach trial-and-error?
32
33	### Missing Context
34	- What information was missing at the start that caused problems later?
35	- What should have been in CLAUDE.md, .cursorrules, or project documentation?
36	- Were there assumptions that turned out wrong?
37
38	### Workflow Improvements
39	- What would make the next session smoother?
40	- Are there recurring patterns that should become rules?
41	- Should any health checks be added to `sheal check`?
42
43	## Step 3: Generate Output
44
45	Present your analysis as a structured retrospective report:
46
47	### Report Format
48
49	```
50	## Session Retrospective: [checkpoint-id]
51
52	### Summary
53	[1-2 sentence summary of what happened and how it went]
54
55	### Health Score: [X/100]
56	[Explain why this score, what went well, what didn't]
57
58	### Key Findings
59
60	#### What Went Well
61	- [thing 1]
62	- [thing 2]
63
64	#### What Could Be Improved
65	- [issue 1]: [why it happened] → [what to do differently]
66	- [issue 2]: [why it happened] → [what to do differently]
67
68	### Suggested Rules
69	[List specific rules that should be added to agent config files]
70	```
71
72	## Step 4: Apply Learnings (with user confirmation)
73
74	After presenting the report, ask the user if they want to apply any of the suggested rules. If yes:
75
76	1. For Claude Code: append rules to `CLAUDE.md` in the project root
77	2. For Cursor: append to `.cursorrules`
78	3. For Gemini: append to `.gemini/GEMINI.md`
79	4. For general: append to `AGENTS.md`
80
81	Format rules as clear, actionable instructions. For example:
82
83	```markdown
84	## Learnings from session [date]
85
86	- Before writing parsers for external data formats, inspect real data samples first (`git show branch:path/to/file \| head -20`). Don't assume formats from documentation alone.
87	- When a CLI tool fails with "connection refused", check if its required server process is running before retrying the same command.
88	- Run `sheal check` at session start to catch environment issues early.
89	```
90
91	## Important Notes
92
93	- Be honest about what went well — not every session has problems
94	- Focus on actionable, specific learnings, not generic advice
95	- Rules should be phrased as instructions to a future AI agent
96	- Don't generate rules for one-off issues that won't recur
97	- Consider whether the issue is project-specific or universal

People who liked this also viewed

6 picks

Type