para-memory-files

Name: para-memory-files
Author: Undertone0809

247stars30forks

How to install para-memory-files?

undertone0809/rudder/para-memory-files

$npx -y skills add Undertone0809/rudder --skill para-memory-files

Installs into the current project.

Summary

File-based memory system using Tiago Forte's PARA method. Use this skill whenever you need to store, retrieve, update, or organize knowledge across sessions. Covers three memory layers: (1) Knowledge graph in PARA folders with atomic YAML facts, (2) Daily notes as raw timeline, (

Files · 1

View on GitHub

SKILL.md

1	# PARA Memory Files
2
3	Persistent, file-based memory organized by Tiago Forte's PARA method. Three layers: a knowledge graph, daily notes, and tacit knowledge. All paths are relative to `$AGENT_HOME`.
4
5	## Three Memory Layers
6
7	### Layer 1: Knowledge Graph (`$AGENT_HOME/life/` -- PARA)
8
9	Entity-based storage. Each entity gets a folder with two tiers:
10
11	1. `summary.md` -- quick context, load first.
12	2. `items.yaml` -- atomic facts, load on demand.
13
14	```text
15	$AGENT_HOME/life/
16	projects/ # Active work with clear goals/deadlines
17	<name>/
18	summary.md
19	items.yaml
20	areas/ # Ongoing responsibilities, no end date
21	people/<name>/
22	companies/<name>/
23	resources/ # Reference material, topics of interest
24	<topic>/
25	archives/ # Inactive items from the other three
26	index.md
27	```
28
29	PARA rules:
30
31	- Projects -- active work with a goal or deadline. Move to archives when complete.
32	- Areas -- ongoing (people, companies, responsibilities). No end date.
33	- Resources -- reference material, topics of interest.
34	- Archives -- inactive items from any category.
35
36	Fact rules:
37
38	- Save durable facts immediately to `items.yaml`.
39	- Weekly: rewrite `summary.md` from active facts.
40	- Never delete facts. Supersede instead (`status: superseded`, add `superseded_by`).
41	- When an entity goes inactive, move its folder to `$AGENT_HOME/life/archives/`.
42
43	When to create an entity:
44
45	- Mentioned 3+ times, OR
46	- Direct relationship to the user (family, coworker, partner, client), OR
47	- Significant project or company in the user's life.
48	- Otherwise, note it in daily notes.
49
50	For the atomic fact YAML schema and memory decay rules, see [references/schemas.md](references/schemas.md).
51
52	### Layer 2: Daily Notes (`$AGENT_HOME/memory/YYYY-MM-DD.md`)
53
54	Raw timeline of events -- the "when" layer.
55
56	- Write continuously during conversations and agent work when a
57	memory-worthy signal appears.
58	- Extract durable facts to Layer 1 during heartbeats.
59	- Treat daily notes as the first capture layer, not the final destination for
60	stable preferences, entity facts, or shared project knowledge.
61
62	#### Conversation and Agent Work Capture Policy
63
64	Rudder chat conversations and agent execution events are memory sources when
65	they contain durable signal, not because every chat line or routine action
66	deserves retention. Capture a concise daily-note entry when a conversation,
67	automation, issue run, review, close-out, or other agent work includes any of
68	these:
69
70	- User corrections that change how the agent should behave next time.
71	- New or changed preferences, constraints, boundaries, or decisions.
72	- Issue proposal intent, acceptance criteria, or priority reasoning that is not
73	already explicit in the issue.
74	- Automation design rationale, recurring workflow choices, or escalation rules.
75	- Agent work patterns that should be repeatable, such as a better triage path,
76	validation shortcut, environment workaround, review handoff, or runbook step.
77	- Project/product/engineering judgment that will affect future work.
78	- Attachment or screenshot evidence that changes task interpretation.
79	- Reusable execution lessons, setup friction, validation findings, or failure
80	modes that future runs should know.
81
82	Do not capture:
83
84	- Greetings, thanks, scheduling chatter, or low-signal status updates.
85	- Routine issue comments, heartbeat updates, command outputs, close-out
86	summaries, or automation logs that add no new decision, correction, lesson, or
87	reusable evidence beyond the source artifact.
88	- Full private chat transcripts. Summarize the durable signal instead.
89	- Secrets, tokens, credentials, private keys, session cookies, or auth headers.
90	- One-time sensitive context that is not needed for future work.
91	- Organization-level facts as private personal memory. Record only the routing
92	decision in the daily note, then promote the fact to shared project knowledge.
93	- Speculation, weak inferences, or unverified assumptions as facts.
94
95	Recommended daily-note entry format:
96
97	```md
98	## HH:MM - Memory capture
99
100	- Context: conversation, issue, automation, run, or evidence reference; project;
101	and why this mattered.
102	- User intent: the durable need, correction, preference, or decision.
103	- Conclusion/action: what changed or what was done.
104	- Reusable

People who liked this also viewed

6 picks

Type

para-memory-files

247stars30forks

View on GitHub

How to install para-memory-files?

undertone0809/rudder/para-memory-files

$npx -y skills add Undertone0809/rudder --skill para-memory-files

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# PARA Memory Files
2
3	Persistent, file-based memory organized by Tiago Forte's PARA method. Three layers: a knowledge graph, daily notes, and tacit knowledge. All paths are relative to `$AGENT_HOME`.
4
5	## Three Memory Layers
6
7	### Layer 1: Knowledge Graph (`$AGENT_HOME/life/` -- PARA)
8
9	Entity-based storage. Each entity gets a folder with two tiers:
10
11	1. `summary.md` -- quick context, load first.
12	2. `items.yaml` -- atomic facts, load on demand.
13
14	```text
15	$AGENT_HOME/life/
16	projects/ # Active work with clear goals/deadlines
17	<name>/
18	summary.md
19	items.yaml
20	areas/ # Ongoing responsibilities, no end date
21	people/<name>/
22	companies/<name>/
23	resources/ # Reference material, topics of interest
24	<topic>/
25	archives/ # Inactive items from the other three
26	index.md
27	```
28
29	PARA rules:
30
31	- Projects -- active work with a goal or deadline. Move to archives when complete.
32	- Areas -- ongoing (people, companies, responsibilities). No end date.
33	- Resources -- reference material, topics of interest.
34	- Archives -- inactive items from any category.
35
36	Fact rules:
37
38	- Save durable facts immediately to `items.yaml`.
39	- Weekly: rewrite `summary.md` from active facts.
40	- Never delete facts. Supersede instead (`status: superseded`, add `superseded_by`).
41	- When an entity goes inactive, move its folder to `$AGENT_HOME/life/archives/`.
42
43	When to create an entity:
44
45	- Mentioned 3+ times, OR
46	- Direct relationship to the user (family, coworker, partner, client), OR
47	- Significant project or company in the user's life.
48	- Otherwise, note it in daily notes.
49
50	For the atomic fact YAML schema and memory decay rules, see [references/schemas.md](references/schemas.md).
51
52	### Layer 2: Daily Notes (`$AGENT_HOME/memory/YYYY-MM-DD.md`)
53
54	Raw timeline of events -- the "when" layer.
55
56	- Write continuously during conversations and agent work when a
57	memory-worthy signal appears.
58	- Extract durable facts to Layer 1 during heartbeats.
59	- Treat daily notes as the first capture layer, not the final destination for
60	stable preferences, entity facts, or shared project knowledge.
61
62	#### Conversation and Agent Work Capture Policy
63
64	Rudder chat conversations and agent execution events are memory sources when
65	they contain durable signal, not because every chat line or routine action
66	deserves retention. Capture a concise daily-note entry when a conversation,
67	automation, issue run, review, close-out, or other agent work includes any of
68	these:
69
70	- User corrections that change how the agent should behave next time.
71	- New or changed preferences, constraints, boundaries, or decisions.
72	- Issue proposal intent, acceptance criteria, or priority reasoning that is not
73	already explicit in the issue.
74	- Automation design rationale, recurring workflow choices, or escalation rules.
75	- Agent work patterns that should be repeatable, such as a better triage path,
76	validation shortcut, environment workaround, review handoff, or runbook step.
77	- Project/product/engineering judgment that will affect future work.
78	- Attachment or screenshot evidence that changes task interpretation.
79	- Reusable execution lessons, setup friction, validation findings, or failure
80	modes that future runs should know.
81
82	Do not capture:
83
84	- Greetings, thanks, scheduling chatter, or low-signal status updates.
85	- Routine issue comments, heartbeat updates, command outputs, close-out
86	summaries, or automation logs that add no new decision, correction, lesson, or
87	reusable evidence beyond the source artifact.
88	- Full private chat transcripts. Summarize the durable signal instead.
89	- Secrets, tokens, credentials, private keys, session cookies, or auth headers.
90	- One-time sensitive context that is not needed for future work.
91	- Organization-level facts as private personal memory. Record only the routing
92	decision in the daily note, then promote the fact to shared project knowledge.
93	- Speculation, weak inferences, or unverified assumptions as facts.
94
95	Recommended daily-note entry format:
96
97	```md
98	## HH:MM - Memory capture
99
100	- Context: conversation, issue, automation, run, or evidence reference; project;
101	and why this mattered.
102	- User intent: the durable need, correction, preference, or decision.
103	- Conclusion/action: what changed or what was done.
104	- Reusable

People who liked this also viewed

6 picks

Type